1. Executive Summary: The Structural Reality of SA1
South Australia represents an extreme frontier of variable renewable energy (VRE) penetration. With rooftop solar frequently meeting over 100% of regional operational demand during midday intervals, the South Australian National Electricity Market (SA1) recorded negative Regional Reference Prices (RRP) in 12.6% of all 5-minute settlement intervals over the trailing 12-month period.
For Battery Energy Storage Systems (BESS), negative spot pricing appears to offer straightforward revenue: assets are paid to absorb energy. However, empirical analysis reveals that naive static heuristic charging rules ("charge whenever price < -$10/MWh") consistently compress net operational margins due to round-trip efficiency (RTE) friction and cyclic cell degradation.
2. The Degradation vs. Negative Spread Tradeoff
A standard 2-hour duration lithium iron phosphate (LFP) BESS exhibits an average AC-to-AC round-trip efficiency (RTE) between 84% and 87%. Every MWh of energy absorbed at a negative price must eventually be discharged at a subsequent positive peak to recover economic value.
To justify entering an active charge interval at price $P_{charge} < 0$, the forward expected discharge price $\mathbb{E}[P_{discharge}]$ within the cycling horizon must satisfy: $$\mathbb{E}[P_{discharge}] \cdot \eta_{RTE} - P_{charge} \ge C_{deg} + C_{FOM}$$ Where $\eta_{RTE} \approx 0.85$, $C_{deg}$ represents cell warranty cycling wear ($\sim \$35-\$48/\text{MWh}$ throughput), and $C_{FOM}$ represents variable O&M.
3. The Pre-Dispatch Trap and 5MS Volatility
Under AEMO's Five-Minute Settlement (5MS) regime, 30-minute pre-dispatch forecasts (P30) and 5-minute pre-dispatch (P5) display significant divergence during sudden cloud cover transit across South Australia. Commercial assets relying on rigid heuristic rules get trapped charging into transient negative intervals that flip positive before dispatch completion.
SparkArb's production model addresses this via an Expected Value (EV) bi-layer framework: Layer 1 evaluates real-time arbitrage against contingency reserve markets, while Layer 2 evaluates multi-feature volatility indicators to filter low-probability trap intervals.
4. Verified Live Record (SA1 Sub-Ledger as of 2026-10-17)
Over the 62-day verified live trading period (commencing August 17, 2026, through settlement date October 17, 2026), SparkArb's SA1 model delivered $60,384.83 AUD in net algorithmic margin on standard 5MW/10MWh reference profiles, outperforming passive benchmark baselines by +19.1% ($60,384.83 vs $50,706.15 AUD baseline) with a 98.4% daily win rate in SA1.
Every daily dispatch plan is cryptographically pre-committed prior to market open with a SHA-256 hash. The full settlement ledger is independently verifiable against public AEMO NEMWEB MMS data.