Industry · Renewable Energy

Renewable energy automation for South African IPPs and traders

Half-hourly meter data reconciled to every wheeling credit, curtailed megawatt-hours evidenced before the compensation claim, and forecasts good enough to hold a wheeling contract inside its tolerance band. Built for NERSA's new rules, not around them.

Discuss your project
6 180 MW
of REIPPPP Bid Window 1–4 capacity online from 90 IPPs, as at 31 March 2025
Source: IPP Office, IPPPP Quarterly Report Q4 2024/25
1 580 MW
of extra wind hosting capacity unlocked in the Cape provinces by accepting 4% curtailment
Source: NTCSA, GCCA-2025 Curtailment Update (Oct 2025)
±5%
tolerance for actual wheeled energy against a three-year forecast in Eskom's wheeling policy
Source: Eskom, Wheeling of Energy and Net-billing Policy (Aug 2025)

Why renewable energy automation is now a settlement problem

Renewable energy automation used to mean SCADA: keep the turbines turning and the inverters online. That work still matters, but in South Africa the harder problem has moved downstream. Since the Electricity Regulation Amendment Act came into force on 1 January 2025, more of the value of a megawatt-hour depends on what happens to its data after it leaves the plant.

Look at what a private generator now has to prove. A wheeling contract on Eskom's network starts with a three-year forecast and is measured against it. Wind farms connecting in the Eastern and Western Cape are doing so on the basis that a share of their output will be curtailed, with compensation for that energy paid within a budget. Grid capacity is allocated on readiness, and the queue is published monthly. Each of those is a data flow between parties who do not share systems, and each one ends in money.

That is where we work. We build the pipelines that pull meter, SCADA and market data into one governed platform, the reconciliation that matches half-hourly generation to time-of-use credits on a customer's bill, the forecasting that keeps contracted volumes inside their tolerance, and the exception queues that put a person in front of anything that does not add up. The AI is useful, but most of the return comes from clean interval data and automated reconciliation.

We have not yet published an energy case study, so we will not pretend otherwise. The closest proof is adjacent: automated reconciliation for a financial services company's month-end, and a streaming, lineage-tracked data platform for a neo bank. Settling wheeled energy is the same discipline applied to a different ledger.

Case study: automated reconciliation for a financial services companyOur data engineering practiceSolar installers and rooftop energy: our solar page
How it works

How a wheeled megawatt-hour gets from meter to invoice

This is the shape of most energy data automation we would build for a generator, trader or wheeling buyer. The pipeline does the collecting, checking and matching; the tolerances you set decide what settles on its own; a person handles everything else.

  1. STEP 01

    Interval data lands

    Half-hourly generation meter data, the offtaker's consumption data and plant SCADA land in one store, raw and unchanged, with the source and time of every file recorded.

  2. STEP 02

    Validated, not silently estimated

    Gaps, duplicates, stuck readings and clock drift are flagged. Missing intervals are marked as missing, so nobody bills on an estimate without knowing it is one.

  3. STEP 03

    Allocated per time-of-use period

    Generation is allocated across offtakers by contract and time-of-use period, checked against the maximum export capacity and against each customer's metered consumption.

  4. STEP 04

    Reconciled against the bill and the forecast

    The allocation is compared with the credits on the network operator's bill and with the contracted forecast, so a variance shows up in days rather than at year-end.

Automated

Within tolerance: invoiced automatically

Where allocation, bill credits and forecast agree within the limits you set, offtaker invoices and statements are generated without anyone re-keying a number.

Human decides

Out of tolerance: routed to an analyst

A missing credit, an export above the maximum export capacity or a forecast drifting outside its band goes to a person with the intervals, the contract clause and the likely cause attached.

Every number traceable

Which meter file, which validation rule and which contract version produced each line on an invoice, retrievable by query when an offtaker, a lender or the regulator asks.

Guides

Renewable energy automation, explained: how to do it and why it is hard

The questions energy teams actually ask us, answered plainly. Each one starts with a short answer you can take into a meeting, then the steps or reasons behind it.

01 · How to

How do you automate energy data flows for wheeling and settlement?

Short answer

Land every half-hourly meter file unchanged, validate it, allocate generation to each offtaker by time-of-use period, then reconcile that allocation against the network operator's bill credits. Automate the matches within tolerance and route the exceptions to a person. Eskom measures IPP output half-hourly but reconciles monthly by time-of-use period.

  1. 1

    List every data source: generation meters, offtaker meters, SCADA, the network operator's bills and the contracts themselves.

  2. 2

    Land raw files unchanged with their source and arrival time, so any number can be traced back later.

  3. 3

    Validate before you allocate: flag gaps, duplicates, stuck values and clock drift instead of quietly filling them.

  4. 4

    Encode each contract's allocation rules, including maximum export capacity and the rule that wheeled energy cannot exceed a customer's metered consumption.

  5. 5

    Reconcile allocation against bill credits every period, and set tolerances with your commercial team, not your vendor.

  6. 6

    Send exceptions to an analyst with the intervals and contract clause attached, and keep a log of how each one was resolved.

Source: Eskom, Wheeling of Energy and Net-billing Policy (Aug 2025)
02 · How to

How does AI improve renewable energy forecasting for wheeling contracts?

Short answer

By re-forecasting as weather and plant availability change, instead of relying on a yield estimate from the financial model. Eskom's wheeling policy asks generator and load customer for a three-year forecast and sets a ±5% tolerance on actual wheeled volumes, so forecast quality now carries commercial weight, even before penalties apply.

  1. 1

    Start from clean history: interval generation, availability and curtailment flags, so the model learns the plant rather than its outages.

  2. 2

    Add weather inputs that match the horizon you need, from next-day to seasonal.

  3. 3

    Forecast at the level the contract is measured at, whether that is a plant, a portfolio or a customer's time-of-use periods.

  4. 4

    Compare forecasts with contracted volumes continuously and flag drift early.

  5. 5

    Keep a simple baseline running alongside the model, so you can show the model is actually earning its keep.

Source: Eskom, Wheeling of Energy and Net-billing Policy (Aug 2025)
03 · Why

Why is curtailment becoming a data problem for South African wind farms?

Short answer

Because new Cape connections are built around it. NTCSA found the Eastern and Western Cape grid fully allocated, and NERSA approved congestion curtailment so more wind could connect, with compensation for curtailed energy paid within a budget based on 4% curtailment. Being compensated fairly means proving, interval by interval, what you could have generated.

  • Curtailment is now planned, not exceptional: NTCSA says it lifts Cape wind hosting capacity from 5 625 MW to 7 205 MW by 2028.

  • The approval runs from 1 April 2025 to 31 March 2028, and compensation is limited to the budget allocated in NTCSA's MYPD6 determination.

  • Lost energy has to be estimated from wind, availability and plant data, which only works if those data are clean and time-aligned. (This point and the next are what we see in the work, not findings from the document.)

  • The plant, the system operator and lenders each hold part of the record, so the evidence has to survive comparison with someone else's numbers.

Source: NTCSA, GCCA-2025 Curtailment Update (Oct 2025)
04 · Why

Why is grid access so hard to get in South Africa, and what data helps?

Short answer

Because the best wind and solar areas are already full. NTCSA reports Cape grid capacity fully allocated through earlier bid windows and private offtakers. NERSA's Grid Capacity Allocation Rules, gazetted in December 2025, now allocate on demonstrated readiness and require network operators to publish the queue monthly, which rewards developers who track their evidence as data.

  • Capacity follows readiness: regulatory, technical, financial and contractual conditions, verified by the network operator.

  • Capacity can be revoked for missing project milestones, so milestone tracking is not an admin nicety.

  • Monthly queue reports include capacity applied for, technology, location, status, position and queued capacity per node.

  • Combining published queue data with hosting-capacity assessments and your own milestones gives a clearer view of which projects can actually connect.

Source: NERSA Grid Capacity Allocation Rules, Government Gazette 53914 (24 Dec 2025)
05 · How to

How should a renewable energy company start with AI automation?

Short answer

Start with one data flow that ends in money, such as wheeling reconciliation or curtailment evidence, and measure how long it takes and how often it is wrong today. Fix the interval data first, automate the reconciliation, and add models only where they beat a simple baseline. Platform programmes can wait until one pipeline has paid back.

  1. 1

    Pick a process with a clear before-number: days to reconcile, invoice disputes, or curtailed megawatt-hours unclaimed.

  2. 2

    Get read access to the meter, SCADA and billing data before anyone writes a model.

  3. 3

    Build validation and lineage first, because every later step depends on trusting the intervals.

  4. 4

    Automate the deterministic work, allocation and reconciliation, before the predictive work.

  5. 5

    Add forecasting or anomaly detection once the data is clean and a baseline exists to beat.

  6. 6

    Keep a person accountable for anything that changes an invoice, a claim or a regulatory return.

Challenges We Solve

The hard problems Renewable Energy faces

Curtailment You Cannot Prove

Connecting in the Cape now comes with planned congestion curtailment and compensation within a budget. Without interval-level evidence of what the plant could have produced, curtailed energy becomes an argument rather than a claim.

Wheeling Reconciliation by Spreadsheet

Half-hourly generation, monthly time-of-use credits, several offtakers and an Eskom or municipal bill that arrives weeks later. Done by hand, the errors are found when an offtaker queries an invoice.

Forecasts That Carry Penalties

Wheeling agreements are measured against contracted forecasts, and Eskom's policy sets a ±5% band. A forecast built once in a financial model and never updated is a commercial risk, not just an engineering one.

Telemetry Spread Across OEM Systems

Turbine, inverter, battery and substation data sit in different vendors' SCADA and historians, with different tags and clocks. Before any model can help, the data has to agree on what time it is.

By line of business

Where automation fits, by type of energy business

Renewable energy is not one buyer. An IPP with a government PPA, a developer in the grid queue, a trader and a municipality all have data problems, and they are different ones.

First win: Performance and reporting

IPPs and asset owners

Operating plants generate continuous SCADA, meter and maintenance data across several OEM systems. Pulling it into one model turns availability, curtailment and lender reporting into a scheduled run instead of a month-end spreadsheet exercise.

First win: Grid access evidence

Developers in the grid queue

Capacity is now allocated on readiness, and network operators publish the queue monthly. Tracking hosting capacity, queue position and milestone evidence as data keeps a portfolio decision grounded in what the grid can actually take.

First win: Settlement and reconciliation

Traders and C&I wheeling buyers

Wheeling pairs half-hourly generation with time-of-use credits on someone else's bill. Portfolio deals multiply that across generators and load accounts, so automated reconciliation is the difference between a margin and a dispute.

First win: Meter data and wheeling admin

Municipalities and distributors

Distributors that allow wheeling inherit the metering, validation and billing work, often on systems never designed for it. Governed meter data pipelines make that administration repeatable and keep customer data handled with consent.

Our Solutions

What we build for you

01

Energy Data Platform

A governed time-series platform for meter, SCADA, weather and market data, with validation, lineage and access control built in, on Azure, Google Cloud or AWS.

02

Forecasting & Curtailment Analytics

Weather-informed generation forecasts that are re-run as conditions change, and lost-energy models that estimate what a curtailed plant would have produced, interval by interval.

03

Settlement & Reporting Automation

Wheeling allocation, bill-credit reconciliation, offtaker invoicing and regulatory reporting generated from tested tables, with exceptions routed to a person rather than buried in a workbook.

Regulatory context

The rules now shaping renewable energy data in South Africa

The market is being rebuilt in public, and most of the documents are available to read. Here is where the main instruments stood as at October 2026, with dates, so you can check rather than take our word for it.

  1. In force since 1 January 2025

    Electricity Regulation Amendment Act 38 of 2024

    Signed in August 2024 and brought into operation on 1 January 2025, the Act provides for a licensed market operator running a transparent, non-discriminatory trading platform and a market code. It is the legal basis for most of what follows.

  2. NERSA approval 29 April 2025, valid to 31 March 2028

    Congestion curtailment in the Eastern and Western Cape

    NERSA designated congestion curtailment as a constrained generation ancillary service. Compensation for curtailed energy is paid within a budget allocated to NTCSA based on a 4% curtailment level, which NTCSA says lifts wind hosting capacity in the two provinces from 5 625 MW to 7 205 MW by 2028.

  3. Published August 2025

    Eskom Wheeling of Energy and Net-billing Policy

    Requires a three-year forecast of wheeled energy, sets a ±5% tolerance (the penalty awaits NERSA approval), measures IPP output half-hourly and reconciles monthly by time-of-use period. Portfolio wheeling is capped at 10 generators and 50 load accounts per portfolio; larger deals go to virtual wheeling.

  4. Approved 12 November 2025, gazetted 24 December 2025

    Grid Capacity Allocation Rules

    Allocates scarce grid capacity on a "first ready" basis rather than first come, and requires network operators to publish a queue status report monthly, including queued capacity per node and estimated timelines. Readiness has to be demonstrated, which makes it a documentation and data exercise.

  5. Licence approved November 2025; launch planned for Q3 2026, go-live not yet confirmed

    Market operator licence and SAWEM

    NERSA approved NTCSA's market operator licence in November 2025. In March 2026 NTCSA moved the wholesale market launch from 1 April 2026 to the third quarter of 2026. We have not been able to confirm independently whether the market has gone live, so check before planning around it.

  6. Revised draft out for comment from June 2026 — not yet final

    NERSA Rules for Electricity Trading

    The drafts envisage real-time meter data shared with sellers where the customer consents, reconciliation by time-of-use period, and semi-annual reporting by traders to NERSA. The detail may still change, but the direction is clear: trading runs on interval data, and that data has to be auditable.

  7. Final draft approved by Cabinet, 15 October 2025

    Integrated Resource Plan 2025

    Cabinet described the plan as a R2.23 trillion investment that will define South Africa's energy mix, with a goal of a net-zero electricity sector by 2050. More private generation means more contracts, more meters and more reconciliation.

Where we start

Engagements we run in Renewable Energy

01

Wheeling reconciliation and offtaker invoicing

Half-hourly generation allocated to each offtaker by time-of-use period, matched to the credits on the network operator's bill, and turned into invoices, with mismatches queued for an analyst instead of discovered in a dispute.

02

Curtailment and lost-energy evidence

Models that estimate what a curtailed plant would have produced from wind or irradiance, availability and plant data, so a compensation claim rests on interval-level evidence rather than an end-of-month estimate.

03

Forecasts that keep contracts in tolerance

Weather-informed generation forecasts re-run as conditions change and compared with contracted volumes, so a wheeling portfolio drifting towards the edge of its band is flagged while there is still time to act.

04

One model across OEM SCADA systems

Turbine, inverter, battery and substation telemetry from different vendors mapped to a common asset model and clock, so performance, availability and fault analysis work across the whole portfolio, not one vendor at a time.

05

Grid queue and capacity tracking

Published hosting-capacity assessments and monthly queue reports pulled into a single view against a developer's own project milestones, so pipeline decisions reflect what the grid can actually take.

06

Lender, offtaker and regulatory reporting

Availability, generation, curtailment and settlement reports generated from the same tested tables every period, so the numbers in a lender report, an offtaker statement and a trader's NERSA return agree with each other.

Proof

Work we have shipped in regulated financial services

Same regulators' expectations, same compliance shape: automate the clear cases, route the uncertain ones to a person, evidence every decision.

Automating Accounting Data Integration for a South African Financial Services Company
Case study · Finance

Automating Accounting Data Integration for a South African Financial Services Company

8 days
Month-end close reduced from 8 days to 2
100%
Automated reconciliation vs. manual process
Read the case study
Cloud Data Architecture for a South African Neo Bank on Google Cloud
Case study · Banking

Cloud Data Architecture for a South African Neo Bank on Google Cloud

10×
Data processing throughput improvement
99.9%
Data quality SLA achieved at gold layer
Read the case study
Key terms

Renewable Energy terms, in plain language

IPP (independent power producer)
A privately owned generator that sells electricity under a power purchase agreement, to Eskom through a government programme or directly to private buyers.
REIPPPP
The Renewable Energy Independent Power Producer Procurement Programme, run by the IPP Office, through which government has procured utility-scale renewable generation in competitive bid windows.
Wheeling
Moving electricity from a generator to a buyer over a network the buyer and generator do not own, with the network operator crediting the buyer's bill for the energy delivered.
Portfolio and virtual wheeling
Wheeling from several generators to several load customers. Eskom caps a portfolio at 10 generators and 50 load accounts; larger deals use virtual wheeling, which currently has no such limit.
Congestion curtailment
The controlled reduction of a generator's output when transmission limits are reached, used in the Eastern and Western Cape to connect more wind than the grid could otherwise host.
GCCA (Generation Connection Capacity Assessment)
NTCSA's published assessment of how much generation the transmission network can accept at each substation and supply area.
Time-of-use (TOU) period
The peak, standard and off-peak blocks a tariff is priced in. Wheeling credits are matched to the period in which the energy was generated and consumed.
SAWEM
The South African Wholesale Electricity Market, to be run by NTCSA as licensed market operator under the amended Electricity Regulation Act.
FAQ

Renewable Energy questions

From our blog

Resilient data pipelines: building for South Africa's realitiesEngineering data management: beyond the spreadsheet5 data quality pitfalls silently killing your automation

Sources & further reading

Related Services
Cloud Data PlatformsDashboards & ReportingAI Automation

Ready to transform your Renewable Energy data strategy?

Let's discuss the specific challenges your business faces and design a solution that delivers real impact.

Book a free AI assessmentView our services