Wheeling is often described as moving electricity from a generator to a distant buyer through the grid. For financial modelling, the more useful view is an accounting and settlement chain. Eskom describes wheeling as balancing generator energy with end-user consumption within time-of-use periods rather than tracing specific electrons from one point to another.
That distinction changes how a corporate PPA model should be built. The model needs three linked views: what the generator produces and invoices, how the network allocates and charges for that energy, and what the buyer pays after the wheeled energy is credited against its normal supply position.
1. Start with hourly or time-of-use generation and load
An annual MWh comparison is not enough. A solar generator may produce a large volume while the buyer's load profile creates limited value in the relevant time periods. Where settlement uses time-of-use periods, generation and consumption should be allocated to those same periods before calculating credits and savings.
This lets the model distinguish three quantities: generated energy, energy that can be matched or allocated to the buyer, and residual grid consumption that remains after wheeling.
2. Keep the PPA price separate from the network settlement
The generator and buyer agree the commercial PPA price. Eskom states that it does not get involved in the PPA between buyer and seller. The network arrangements sit alongside that bilateral contract.
That means the buyer's economic benefit cannot be measured by simply subtracting the PPA tariff from an Eskom tariff. The model should separately calculate the PPA payment, the normal utility bill, the wheeling credit or adjustment, losses, applicable network or administration charges and any residual energy purchased from the utility.
3. Model losses explicitly
Network losses affect how much economic value reaches the buyer. Eskom's current wheeling explanation notes that buyer credits are based on wheeled energy within the relevant TOU periods and account for energy losses. A model that assumes one generated MWh produces one full MWh of buyer credit can therefore overstate savings.
4. Residual supply is part of the corporate power cost
Most buyers remain dependent on the grid for periods when the renewable generator does not match consumption. The model should therefore keep the residual utility bill visible. This is especially important for solar because the buyer's load may continue through evening and off-peak periods.
The correct comparison is the buyer's total electricity cost before the transaction versus the total delivered cost after the transaction, not the PPA invoice alone.
5. Network and administration charges should be treated as real project economics
Eskom's wheeling framework distinguishes generator use-of-system charges, connection charges and administration associated with wheeling. The exact tariff schedules change over time, so a live model should use the current approved schedules and the buyer's actual tariff class.
Rather than hard-coding one generic "wheeling fee", keep the charge categories visible. That makes it easier to update the model when tariffs change and to see which part of the economics is driven by the network rather than the PPA.
6. Contract structure determines whose cash flow is exposed
The generator may face curtailment, connection risk, imbalance or settlement timing risk. The buyer may face volume mismatch, residual grid price increases, credit adjustments and early-termination exposure. A finance model should keep these risks with the party that actually bears them under the PPA and network agreements.
If a generator's debt model assumes full contracted revenue but the PPA only pays for allocated or delivered energy, the revenue case needs to be reconciled with the wheeling settlement mechanics.
7. Buyer savings and generator returns need to work at the same time
A corporate PPA is commercially durable only if both sides can live with the economics. The buyer wants lower and more predictable delivered power cost. The generator needs enough contracted revenue to support operating costs, debt service and investor returns.
The model should therefore show both views together. A tariff that creates attractive buyer savings may still be too low to finance the project, while a tariff that produces strong project returns may offer insufficient value to the buyer once network effects are included.
A practical wheeling model should answer
- How much generation is matched to buyer load in each TOU period?
- What is the PPA payment to the generator?
- What losses and network charges apply?
- What residual utility energy does the buyer still purchase?
- What is the buyer's total delivered electricity cost before and after the transaction?
- How do tariff escalation, generation downside and network-charge changes affect savings?
- Does the generator still meet debt-service and return requirements under the same downside cases?
Current-framework note: South African wheeling arrangements and tariff schedules continue to evolve. For a live transaction, confirm the current Eskom, NERSA and relevant municipal requirements rather than relying on a static historical assumption set.