Can you deliver the next three hours?
A site needs 20 MW less grid import through the evening. Work out whether stored energy and a temporary reduction in customer demand can meet the commitment—and be ready when called again.
Can the battery cover the peak?
Supply three one-hour demand blocks without exceeding a 40 MW grid-import limit. Adjust battery power and starting energy.
Dashed line: grid-import limit. Brown is demand this plan cannot supply within the limit—not a forecast blackout.
Energy is already usable AC-delivered MWh. No recharge during this event; no additional loss deduction.
The simplified schedule meets the import limit.
Check the hourly dispatch
| Hour | Demand MW | Grid MW | Battery MW | Gap MW | Energy left MWh |
|---|---|---|---|---|---|
| 1 | 20 | 20 | 0 | 0 | 30 |
| 2 | 45 | 40 | 5 | 0 | 25 |
| 3 | 65 | 40 | 25 | 0 | 0 |
For this exact sequence, the minimum is 25 MW and 30 MWh. Dispatch each hour is limited by required reduction, available MW and remaining MWh ÷ hours. This is a feasibility check, not a price-optimising dispatch model.
Original synthetic case. Before using a schedule at work, check ramping, state-of-charge limits, auxiliaries, temperature, reserve obligations, outage risk and recovery for the next event.
The lesson
10 chaptersBrowse the chapters
01Specify the service before choosing the asset
The evening commitment
A site must reduce its grid import by 20 MW from 18:00 to 21:00. It can use a battery behind the meter, reduce some production temporarily, or combine both. A battery sales sheet says 30 MW. Is that enough? You need a duration, an energy boundary and a recovery plan before you can answer.
Turn the request into something measurable
- 01
Where and when?
Specify the connection meter, start time, notice period and response deadline. Flexibility in another location may not relieve this constraint.
- 02
How much and how long?
State the reduction in MW and the event length. Here, 20 MW for three hours means 60 MWh of import reduction.
- 03
What happens afterwards?
State how the store recharges or production catches up, and when another call must be possible.
| Need | Measure | Do not assume |
|---|---|---|
| Respond quickly | Delay and ramp rate | Fast controls guarantee sufficient stored energy |
| Sustain delivery | MW across the whole event | A nameplate MWh rating is all available now |
| Be ready again | Recovery time and end inventory | The asset starts full again tomorrow |
Check your understanding
A 30 MW / 40 MWh usable-delivered battery is offered for the 20 MW, three-hour commitment. Is it sufficient alone?
Show reasoning
No. It has enough power, but only two hours of energy at 20 MW.
40 MWh ÷ 20 MW = 2 h. The commitment requires 60 MWh, so it is 20 MWh short before considering other limits.
02Find the energy that can actually be delivered
From internal inventory to AC delivery
Edelivered = (Estart − Emin) × ηdischarge
Constant discharge efficiency; no separately modelled auxiliary load, self-discharge or temperature derating. Do not apply this deduction again to an already usable AC-delivered rating.
A 200 MWh rating is not 200 MWh available
A nominal four-hour battery
Identify the operating window
200 × (0.90 − 0.10) = 160 MWh internally
The nameplate is 200 MWh, but the permitted range is only 80% of it.
Convert to delivered energy
160 × 0.95 = 152 MWh AC
This is the energy available at the specified delivery boundary under the stated assumptions.
Test a 50 MW request
152 ÷ 50 = 3.04 hours
The nameplate ratio 200/50 = 4 hours does not describe this usable operating window.
Result: Ask for initial inventory, reserve floor, delivery boundary and auxiliary treatment—not just the brochure's duration.
Check your understanding
The same 200 MWh store begins at 50% rather than 90%, with the same 10% floor and 95% discharge efficiency. How much can it deliver?
Show reasoning
76 MWh AC.
200 × (0.50 − 0.10) × 0.95 = 76 MWh. At 50 MW it lasts 1.52 hours, not 3.04.
03Keep a chronological inventory ledger
One interval at a time
Enext = Enow + Pcharge × Δt × ηcharge − Pdischarge × Δt ÷ ηdischarge
MW at the AC boundary and MWh internally. No simultaneous charge and discharge in this model. Enforce minimum/maximum energy and power limits in every interval.
| Step | AC transfer | Internal change | End inventory |
|---|---|---|---|
| Start | — | — | 20 MWh |
| Charge for 1 h; η = 0.90 | 40 MWh in | +36 MWh | 56 MWh |
| Discharge for 1 h; η = 0.90 | 20 MWh out | −22.22 MWh | 33.78 MWh (rounded) |
Why division appears on discharge
To deliver 20 MWh after a 90% conversion, the store must release 20/0.90 = 22.22 MWh internally. Multiplying by 0.90 instead would remove too little energy and overstate the next interval's availability. Keep full precision in your model and round only the displayed results.
Check your understanding
A store needs 72 MWh of internal replenishment, with 20 MW charging power and 90% charging efficiency. Minimum constant-power charging time?
Show reasoning
4 hours.
Each hour adds 20 × 0.90 = 18 MWh internally. 72/18 = 4 h, excluding interruptions, tapering and other loads.
04Demand response: verify the reduction and the rebound
The meter shows what happened—not what would have happened
To estimate a load reduction, compare actual demand with a justified no-response baseline. Production, weather, occupancy and maintenance can change that baseline. A factory closure that was going to happen anyway should not be presented as a newly delivered response.
| Period | Change against justified baseline | Duration | Energy effect |
|---|---|---|---|
| Event | 4 MW less import | 2 hours | 8 MWh avoided |
| Recovery | 3 MW more import | 2 hours | 6 MWh added |
| Both periods | — | 4 hours | 2 MWh net reduction |
Separate event benefit from whole-window energy
Net energy reduction = event reduction − recovery energy
Use a consistent counterfactual and complete recovery window. This simplified calculation is not a wholesale settlement methodology.
A useful shift need not save much energy
Moving refrigeration or pumping away from a constrained peak can help even if the deferred work later consumes the same energy. But rebound during another constrained period can move the problem rather than solve it. Include production quality, comfort, worker safety and restart limits in the service agreement.
Check your understanding
A site claimed 4 MW reduction from a 12 MW baseline, but an independently justified baseline is 10 MW. Actual demand was 8 MW for two hours. What reduction follows?
Show reasoning
2 MW, or 4 MWh during the event.
(10 − 8) × 2 = 4 MWh, half the earlier 8 MWh claim. The result is sensitive to the counterfactual, not just meter accuracy.
05Compare flexibility technologies by their limits
| Resource | Energy or recovery constraint | Evidence needed |
|---|---|---|
| Battery | Initial inventory, usable range, efficiency and degradation | AC power/energy envelope, auxiliary load and warranty assumptions |
| Pumped hydro | Water inventory, head, pumping schedule and water constraints | Reservoir operating range, generation/pumping curves and connection studies |
| Demand response | Customer process, baseline, notice and rebound | Verified delivery record, baseline method and recovery plan |
| Flexible generation | Fuel access, starts, minimum loading and ramp limits | Fuel/transport availability, operating envelope and emissions obligations |
Storage shifts energy; it does not supply its own recharge
Pumped hydro uses electricity to lift water and later generates as water flows down through turbines. Batteries also require a charging source. A storage proposal therefore needs both a discharge plan and feasible replenishment. A renewable annual-energy surplus does not establish that charging is available at the right place and time.
A simple full-cycle efficiency
ηround-trip = EAC,out ÷ EAC,in
Compare a cycle returning to the same internal inventory at the same boundary. Without separately modelled losses, ηround-trip = ηcharge × ηdischarge.
Check your understanding
Both charging and discharging efficiencies are 90%. Is round-trip efficiency 90%?
Show reasoning
No. In this model it is 81%.
100 MWh AC in becomes 90 MWh internally and then 81 MWh AC out. Auxiliary consumption or other losses can lower the measured result further.
06Cover the three-hour evening commitment
Use these inputs—not a manufacturer's specification
Return to the site's 20 MW reduction for three hours. The teaching battery starts with 90 MWh internally, has a 10 MWh floor, a 100 MWh ceiling and a 30 MW discharge limit. Assume 80% discharge efficiency, 90% charging efficiency and no other losses. A customer can reduce its underlying load by 5 MW during the first two hours only; that reduction is additional to battery delivery, so there is no double counting.
| Hour | Import reduction required | Customer response | Battery output | Internal energy used |
|---|---|---|---|---|
| 18:00–19:00 | 20 MW | 5 MW | 15 MW | 18.75 MWh |
| 19:00–20:00 | 20 MW | 5 MW | 15 MW | 18.75 MWh |
| 20:00–21:00 | 20 MW | 0 MW | 20 MW | 25 MWh |
Check feasibility before price
Power in every hour
Battery output: 15, 15 and 20 MW
All are within the 30 MW power limit. The first two hours also require the promised 5 MW customer reduction.
Total energy
Battery: 50 MWh AC; customer response: 10 MWh
Together they provide the required 60 MWh of import reduction over the event.
End inventory
90 − 50/0.80 = 27.5 MWh internally
The store remains 17.5 MWh above the 10 MWh floor. Check the intermediate states too: 71.25, 52.5 and 27.5 MWh.
Result: The combination meets this event under these assumptions. Battery-only delivery is also feasible here: 60/0.80 = 75 MWh internally, leaving 15 MWh. Comparing cost and recovery remains necessary.
Check your understanding
The customer cannot reduce load in the last hour. Is the combined schedule still feasible?
Show reasoning
Yes—that is already built into the schedule.
Customer response is 5, 5 and 0 MW. Battery output rises to 20 MW in hour three. Assuming a third hour of response would incorrectly save another 5 MWh in the model.
07Test recovery and a second call
The first successful event is not the end of the story
After the combined schedule, the battery contains 27.5 MWh internally. A second identical 20 MW, three-hour call may arrive after a two-hour recovery gap. The customer cannot provide demand response again. Assume the battery may charge at 10 MW throughout that gap and charging efficiency is 90%.
Does the second call fit?
Replenish during the gap
10 MW × 2 h × 0.90 = 18 MWh internally
The 10 MW allowance is explicitly assumed to remain available after allowing for the customer's rebound and other site load.
Find deliverable energy at the next start
(27.5 + 18 − 10) × 0.80 = 28.4 MWh AC
Do not count the reserve floor as usable energy.
Compare with the next obligation
60 − 28.4 = 31.6 MWh short
At 20 MW, the available energy supports 1.42 hours. The 30 MW inverter does not solve the missing energy.
Result: This portfolio can meet the first call but not the second as specified. Add feasible recharge time, another resource or a smaller contractual commitment before promising repeated delivery.
Recovery changes cost as well as readiness
Suppose the customer uses an extra 4 MWh after its 10 MWh event reduction. Its net whole-window energy reduction is 6 MWh. If this rebound overlaps battery charging, assess their combined site import. A recharge assumption that ignores the connection limit is not a feasible recovery plan.
Check your understanding
How much AC charging energy restores the battery from 27.5 to 90 MWh internally at 90% efficiency?
Show reasoning
About 69.44 MWh AC, requiring about 6.94 hours at 10 MW.
Internal replenishment is 62.5 MWh. Divide by 0.90 to obtain charging energy, then divide by 10 MW. This exceeds the two-hour recovery gap.
08Write a service commitment, not a technology slogan
Your screening assignment
Reproduce the three-hour combined schedule, then remove the customer response and reduce the initial battery inventory from 90 to 70 MWh. Keep the 10 MWh floor, 80% discharge efficiency and 30 MW output limit. Calculate the sustained delivery time, missing MWh and what additional evidence would be needed before a real commitment.
| Section | What belongs in it |
|---|---|
| Service | Location, 20 MW requirement, three hours, notice and response deadline |
| Physical ledger | Per-hour battery output, customer reduction and internal inventory |
| Recovery | Recharge power/time, rebound and next-call readiness |
| Commercial comparison | Charging cost, customer compensation, degradation and opportunity cost |
| Evidence and limits | Baseline validity, site import/export capability, outages and operational approval |
Compare cost on the same boundary
A separate synthetic arbitrage cycle takes in 100 MWh at A$30/MWh and returns 81 MWh at A$100/MWh, ending at its starting inventory. Gross energy margin is A$8,100 − A$3,000 = A$5,100, before degradation, fees and other costs. A price spread alone is insufficient: quantity, efficiency and the energy reserved for another service all matter.
Check your understanding
What changes when the battery starts at 70 MWh and customer response is unavailable?
Show reasoning
It can deliver 48 MWh: 2.4 hours at 20 MW, leaving a 12 MWh shortfall.
(70 − 10) × 0.80 = 48 MWh. It passes the power test but fails the energy test. A reviewable recommendation identifies the gap rather than hiding it in an average.
09Diagnose the bottleneck before buying more capacity
| Battery, usable AC basis | Outcome in the interactive case | Reason |
|---|---|---|
| 25 MW / 30 MWh | Meets the three-hour schedule exactly | Uses 5 MWh in hour 2 and 25 MWh in hour 3 |
| 20 MW / 60 MWh | 5 MWh remains uncovered in hour 3 | Power binds even though 35 MWh remains stored |
| 25 MW / 20 MWh | 10 MWh remains uncovered in hour 3 | Only 15 MWh remains after hour 2 |
A single-interval upper bound
Pdelivery ≤ min(Pavailable, Eremaining / Δt)
Remaining energy is on the delivered AC basis used by this exercise. For internal battery energy, account for discharge efficiency and the operating floor first.
Write the diagnosis, not just the score
For 20 MW / 60 MWh, request at least 5 MW more discharge capability or reduce demand in the final hour. For 25 MW / 20 MWh, at least 10 MWh more starting usable energy is needed for this schedule. These are teaching minima, not procurement recommendations with engineering margins.
Check your understanding
Does changing the 20 MW / 60 MWh battery to 20 MW / 100 MWh fix the final hour?
Show reasoning
No. The last hour still requires 25 MW of battery output.
Extra inventory does not increase the 20 MW discharge limit. The plan must change power capability, demand, another supply source or the permitted import limit.
10A feasible evening is not a complete operating plan
Follow the inventory chronologically
Start at 30 MWh
Hour 1: 30 − 0 = 30 MWh
Demand is below the import cap. This no-recharge exercise does not add energy.
Cover the first excess
Hour 2: 30 − 5 = 25 MWh
45 MW demand minus 40 MW grid import leaves 5 MW for one hour.
Cover the peak
Hour 3: 25 − 25 = 0 MWh
The 65 MW demand needs 25 MW from the battery. There is no remaining buffer.
Result: Exactly feasible under the stated assumptions; no spare usable energy at the end.
| Check | Why it matters |
|---|---|
| Initial state of charge and reserve floor | Nameplate energy may not be available at event start |
| Ramps, controls and connection constraints | A steady block model ignores transient limits |
| Outages, temperature and degradation | Available MW and MWh vary with asset condition |
| Recovery and next event | A discharged asset is not ready again without a recharge plan |
| Sensitivity and margin | Exact equality is not resilience to forecast error |
Check your understanding
The first evening ends empty. Can you claim the same response is available the following evening?
Show reasoning
Not without a feasible recovery plan and a verified starting inventory.
Specify when charging occurs, how much grid headroom and charge power are available, the energy boundary and losses, and any competing obligations. Repeating the same starting-energy assumption is not evidence.
Watch the explanation
Key terms in plain English
- MW
- Megawatt: a measure of power, or how quickly energy is transferred. A plant's maximum MW is not its actual output in every interval.
- MWh
- Megawatt-hour: a quantity of energy. A steady 1 MW for one hour delivers 1 MWh.
Supporting diagrams
One complete storage cycle
Declare where each energy quantity is measured.
- 01 · AC input100 MWh
Charging efficiency 90%
- 02 · Stored internally90 MWh
Discharge efficiency 90%
- 03 · AC output81 MWh
Round-trip efficiency 81%