Topic
Actuarial energy modelling: Expert analysis for energy decision-makers
Discovery Green
By Andre Nepgen, Head of Discovery Green 11 June 2026
Securing a long-term energy contract requires certainty against weather variability, business demand fluctuations, and generation asset performance to ensure your budget aligns with actual output. Discovery Green's actuarial methodology underlying EnergyOS treats this requirement the same way actuaries have managed insurance risk for 30 years: using probability distributions, data, and confidence intervals to set contract terms against what generation assets will actually deliver, not what they average. This article explains how the methodology works and why it changes the contract your business signs.
EnergyOS applies statistical methods to renewable energy supply. Rather than estimating annual generation from a single average figure, the approach uses probability distributions and P90 confidence intervals to determine and understand what a generation asset will deliver under any future. Discovery Green applies this to every contract.
Key takeaways:
- P90 confidence intervals: P90 confidence intervals tell you the floor of generation delivery, not the average. The floor is the number that belongs in a board-level risk assessment.
- Interval consumption data: Twelve months of 30-minute interval consumption data, analysed before any contract is set, ensures the contract structure reflects your business's forecasted demand curve, not a flat proxy.
- Portfolio diversification: Portfolio diversification across solar and wind on geographically separate sites means one underperforming asset does not determine your contract's delivery outcome.
- Contract structures: Contract structures span 5 to 20 years with CPI-linked escalation. The actuarial analysis determines which structure and product match your business's risk profile.
Request your actuarial energy analysis
What actuarial energy modelling actually means
EnergyOS's generation forecasting model actuarial and statistical methods to renewable energy supply, using historical weather data, seasonal patterns and the operating characteristics of each renewable energy asset, to forecast how much electricity it is expected to produce and when. Rather than relying on a single average figure or flat-baseline estimate, the model develops a probabilistic generation profile, including P90 confidence intervals to quantity the range of potential outcomes and establish a reliable generation floor. Discovery Green applies this to every contract.
For example, if a generation asset has a P90 of 180 GWh per year, it is expected to produce at least 180 GWh in 9 out of every 10 years. By applying 30 years of Discovery Group actuarial heritage to energy procurement, Discovery Green can account for variability and design every contract around a more reliable view of the renewable energy likely to be available, giving decision makers greater confidence that the figures presented to the board reflect dependable delivery rather than optimistic averages.
This level of analysis is essential when evaluating Power purchase agreement solutions for your corporate facilities.
Why interval consumption data changes the contract
Standard energy quotes often rely on monthly or annual consumption totals. EnergyOS goes deeper, using at least 12 months of half-hourly interval consumption data to build a detailed view of when and how a business consumes energy. This allows Discovery Green to understand the shape of the customer's demand - including daily peaks, seasonal changes and operational patterns - and to model how consumption may respond to structural and operational drivers such as economic growth, seasonality, changes in production and unexpected shock events. By testing a range of plausible demand scenarios, EnergyOS can assess how the customer's future consumption profile may evolve before recommending a product or contract structure. The result is a solution aligned to the business's actual consumption profile, rather than a flat estimate that can obscure the periods of greatest cost or coverage risk.
Demand profiles differ significantly across industries. A three-shift manufacturer or 24/7 mining operation has a very different load shape from a commercial property or hospitality business with pronounced seasonal variability. By analysing these differences alongside forecast renewable generation, EnergyOS can assess how well supply and demand align over time and determine the product, coverage level and contract features best suited to the customer's needs.
This analysis enables Discovery Green to deliver a customised proposal within 10 business days of receiving complete and usable data. Understanding the customer's interval consumption profiles is therefore a foundational step in designing an effective energy-wheeling solution.
How portfolio diversification reduces concentration risk
Discovery Green's 740MW portfolio combines solar and wind across geographically separate sites, reducing reliance on any single technology, asset or weather pattern. A standalone solar or wind asset can expose a customer to concentrated generation risk when local conditions are unfavourable. By modelling the expected output of each asset together, EnergyOS creates a more diversified and resilient view of the renewable energy available to serve customers over time.
This applies the actuarial principle of risk pooling: the generation portfolio is designed so that the performance of one asset or technology does not determine the overall outcome. Discovery Green;s portfolio to date includes the Overberg Wind Farm (150MW, 489 million kWh per year), the Tournee Solar Park (300MW, 720 million kWh per year) and the Notsi Solar Farm (290MW, 914 million kWh per year). Their generation profiles are complementary: solar production is concentrated in daylight hours and is generally stronger in summer, while wind can contribute at different times of day and across different seasons. Energy OS models these profiles together to forecast the combined supply available across the portfolio and manage variability more effectively.
EnergyOS then adds a second layer of diversification by aggregating customers with different consumption profiles and intelligently allocating renewable energy across them. In one modelled example, a business achieving 49% coverage on a standalone basis could reach 78% portfolio coverage when aggregated with 5 businesses from different industries.
This is the structural advantage of platform diversification: it reduces concentration risk at both the generation and customer level, while EnergyOS continuously forecasts, customises and allocates energy to improve each customer's outcome.
Read our methodology: How Discovery Green structures wheeling contracts
Applying actuarial principles to a real contract decision
Discovery Green uses EnergyOS to translate a business's energy data and objectives into a contract structure designed for its needs. The platform brings together three core inputs: the business's historical and forecast consumption profile, the expected generation profile of Discovery Green's diversified renewable-energy portfolio, and the customer's priorities across coverage, savings, price certainty, flexibility and contract duration. EnergyOS models these variables together to recommend a solution that reflects how the business is expected to consume electricity and how renewable energy is likely to be available over time.
The process begins when you submit at least 12 months of half-hourly interval consumption data. EnergyOS uses this information to build and forecast the customer's demand profile, then assesses how it aligns with the expected output of the renewable-energy portfolio. Discovery Green's actuarial team can use this analysis to recommend an appropriate starting product and tailor relevant contract features. These include Green Flex (~30% coverage, a first step into renewable energy solar product), Green Saver (90% coverage, designed to balance high coverage with savings), and Green Guarantee (designed for businesses seeking full renewable-energy coverage). Discovery Green has applied its approach for customers including Impala Platinum Holdings Limited (Implats), Glencore, Fortress Real Estate Investments, Balwin Properties, KP Lime, and The Capital Hotels and Apartments.
Contract options span 5 to 20 years, with CPI-linked escalation. Once complete and usable data has been received, Discovery Green can deliver a customised proposal within 10 business days. Because the modelling traces the proposed coverage, pricing and contract structure back to the customer's own consumption data and defined assumptions, decision-makers can evaluate the commercial and sustainability trade-offs on a transparent basis. The resulting renewable-energy procurement may also support the customer's broader carbon-management strategy, including its response to South Africa's Carbon pricing framework subject to the customer's circumstances and applicable requirements.
Our contracts comply with the National Energy Regulator of South Africa (NERSA) wheeling framework and licensing requirements (nersa.org.za). The Electricity Regulation Act 4 of 2006 (ERA 2006) provides the legal basis for wheeling across the national grid (Government Gazette). Our portfolio expansion aligns with the generation capacity targets set out in the IRP 2023 (Integrated Resource Plan) by the Department of Mineral Resources and Energy.
Once the contract is in place, EnergyOS continues to forecast generation and demand and to allocate available renewable energy across the customer portfolio. This supports transparent reporting of the renewable energy allocated to each customer and enables businesses to assess the associated commercial and Scope 2 emissions benefits in line with the applicable accounting methodology. The result is not simply a standard energy quote, but a data-led solution designed to adapt to the customer's consumption profile and deliver measurable value over the life of the contract.
Request your actuarial energy analysis
Frequently asked questions
What is a P90 confidence interval in the context of energy modelling?
P90 is a probabilistic measure used to estimate a conservative level of renewable energy generation. A P90 forecast indicates that the asset is expected to meet or exceed the stated output in nine out of ten years. For example, if a generation asset has a P90 output of 180 GWh per year, it is expected to generate at least 180 GWh in 90% of years. EnergyOS uses P90 alongside broader generation forecasts, not as a guarantee, but as one confidence measure that helps Discovery Green account for variability and structure contracts around a more reliable view of the renewable energy likely to be available.
Why does Discovery Green require 12 months of interval consumption data before proposing a contract?
At least 12 months of half-hourly interval data gives EnergyOS a detailed view of when and how your business uses electricity across daily, weekly, and seasonal cycles. The platform uses this history to build and forecast your demand profile, assess how it aligns with the expected generation from Discovery Green's renewable-energy portfolio, and test how consumption could change under different scenarios. This enables Discovery Green to recommend an appropriate product, coverage level and contract features based on your operational needs and commercial priorities, rather than relying on monthly or annual averages that can obscure important variations in demand.
Conclusion
Long-term renewable energy procurement requires more than an estimate of annual supply or demand. It requires a reliable view of how renewable generation and business consumption may change over time and a solution designed to perform across those possible outcomes.
EnergyOS combines actuarial modelling, generation and demand forecasting, portfolio diversification and intelligent energy allocation to manage this complexity. By analysing when renewable energy is likely to be available, how a customer is expected to consume electricity and how these profiles align, Discovery Green can structure a solution around the customer's operational needs, commercial priorities and risk appetite.
The result is a more informed approach to renewable energy procurement: one that gives decision-makers a transparent basis for assessing coverage, savings, price certainty and sustainability outcomes over the life of the contract. Request a customised proposal to see how EnergyOS can design a renewable energy solution around your business's demand profile.
About the author
Andre Nepgen is the Chief Executive Officer (CEO) of Discovery Green where leads the development of renewable energy solutions for businesses across South Africa. Drawing on Discovery's actuarial heritage, his work focuses on using data, forecasting and portfolio-based risk management to structure renewable energy contracts that provide greater price certainty, mitigate risk and support measurable environmental progress.
Disclaimer
Discovery Green Registration number: 2023/647514/07. Discovery Limited is the licensed controlling company of the designated Discovery Insurance Group. Registration number: 1999/007789/06. Companies in the Group are licensed insurers and authorised financial services providers. Terms, conditions and limits apply. Discovery Green savings apply to generation costs only. Transmission and distribution charges, levies, and grid charges are unchanged and continue to be billed by the customer's distributor. Individual outcomes, contract terms, and rates depend on consumption profile, product selection, contract length, site eligibility, and grid factors. This information is educational and does not constitute financial, legal, tax, or commercial advice. Always consult qualified energy advisers, legal counsel, and your finance function before signing any wheeling or PPA agreement. Contact: support.discoverygreen@discovery.co.za | +27 11 529 2888 | www.discoverygreen.co.za
Last updated: 11 June 2026