Battery Energy Storage Systems (BESS) are becoming central to NEM stability, arbitrage, and renewables integration. Yet despite rapid deployment, financial outcomes continue to diverge sharply from pre-investment revenue modelling. The common culprit: traditional modelling frameworks break down under high FCAS volatility.
2023–25 has delivered some of the most volatile FCAS conditions since the introduction of the contingency and regulation markets. Price spikes have become less predictable, event durations have become shorter and more clustered, and inter-regional price separation has intensified. Most BESS pro formas—built on smoothing assumptions or simplified dispatch logic—simply cannot capture real-world volatility.
The result? Systematically biased revenue forecasts, mis-optimised bidding behaviour, and an underestimation of portfolio risk.
This article sets out five reasons most BESS revenue models fail under high FCAS volatility, and how advanced modelling—like Northstream Analytic’s PowerStream short-term dispatch engine—overcomes these structural weaknesses.
1. FCAS Volatility Is Not Normally Distributed — But Most Models Pretend It Is
Many investment decks and banker-friendly models still assume:
- Mean-reverting FCAS prices
- Log-normal or normal distributions
- Smooth temporal volatility without clustering
But the NEM’s FCAS markets consistently exhibit:
- Heavy-tailed price distributions
- Clustering of extreme events
- Sudden regime shifts
- State-dependent volatility (e.g., islanding risk, unit outages, wind ramps)
This means that using a single “average” FCAS price per service category is mathematically invalid. The tails drive the earnings: a handful of intervals can deliver >50% of annual FCAS revenue, especially for Raise/Lower Contingency services.
When models smooth this out, they under-forecast revenue in good years and over-forecast in bad years, giving investors a false sense of stability.
2. Static Capacity Allocation Assumptions Break During Price Spikes
A typical investor model sets fixed percentages of battery capacity to:
- Energy arbitrage
- FCAS regulation
- FCAS contingency
But in real operations:
- Energy and FCAS revenue stacks interact dynamically
- The optimal allocation changes every 5 minutes
- FCAS price spikes alter opportunity cost instantly
- State of Charge (SOC) constraints limit which markets are viable
Static partitioning therefore suppresses upside in volatile FCAS conditions.
For example, during a high Raise 6-second event, the optimal strategy may shift from energy discharge to contingency enablement—but this shift is state-dependent and price-dependent, not rule-based.
Models that do not simulate real bidding strategy + SOC dynamics will systematically miss this.
3. SOC Constraints Under Volatility Create Nonlinear Profit Paths
Most simplified revenue models treat SOC as a linear constraint:
“Battery charges at low price, discharges at high price.”
Under FCAS volatility, SOC is instead a multi-dimensional optimisation problem involving:
- Enablement depth
- High-price tail events
- Energy lost to regulation
- Cycling limits and throughput penalties
- Interaction with real-time AGC dispatch
This creates nonlinearities such as:
- Being full too early → missing a late-day FCAS spike
- Being empty during islanding risk → missing Raise contingency
- Misaligned SOC trajectory → forced charging during expensive intervals
These complexities lead to revenue paths that no spreadsheet “averaging logic” can reasonably reproduce.
Only a constraint-based dispatch model can capture these dynamics.
4. FCAS Enablement and Compliance Are Not Perfectly Efficient
Real-world BESS operation loses revenue because:
- Enablement often < 100% due to AGC behaviour
- Causer-pays factors penalise poor performance
- Compliance events reduce available MW
- AGC movement consumes energy and reduces SOC stability
- Thermal and inverter constraints limit maximum contingency enablement
Most revenue models assume perfect availability across all FCAS services.
This is rarely true. In fact, measured availability for some NEM assets can be as low as 70–80% after accounting for compliance and physical derating.
Ignoring this leads to:
- Overestimated FCAS revenue
- Underestimated degradation cycles
- Incorrect valuation of control system quality
These errors compound under volatility.
5. Volatility Interacts With Degradation — A Hidden Cost
High FCAS volatility encourages batteries to chase short, high-value events.
But this often increases:
- Cycling depth
- Micro-cycling from AGC
- C-rate stress
- Temperature-related degradation
Degradation is nonlinear, not proportional to MWh throughput. One week of extreme FCAS activity can generate disproportionately large degradation events.
Simplified models typically assume:
- A constant $/MWh degradation cost
- Even cycling over time
- No interaction between volatility and cell temperature
This is incorrect.
As FCAS volatility grows, ignoring nonlinear degradation overstates revenue and understates O&M reserves.
What a Valid BESS Revenue Model Must Include in a High-Volatility FCAS World
A robust model must integrate:
1. Five-minute NEMDE-aligned price simulation
Including scenario-based volatility and tail-event structure.
2. Full SOC trajectory modelling
Respecting charge/discharge rates, ramping, and energy constraints.
3. Co-optimised energy + FCAS bidding
Not preset allocations.
4. Compliance, enablement and AGC realism
Using empirical derating factors or control-loop simulation.
5. Nonlinear degradation modelling
Based on depth-of-cycle and stress factors.
6. Locational constraints and network constraints
Including MLF variability, curtailment risk, and islanding probabilities.
Conclusion: FCAS Volatility Rewards the Prepared and Punishes the Simplistic
As the NEM becomes more dynamic and renewable penetration increases, FCAS volatility is structurally rising—not falling. Batteries positioned to respond intelligently to this volatility will outperform their peers. But that requires models that reflect market reality, not spreadsheet mythology.
Most BESS investment models fail because they smooth what should be spiky, simplify what is nonlinear, and ignore what is stochastic.
If developers, investors, and operators want accurate forecasts, they need modelling frameworks built for volatility, not stability.
Northstream Analytic’s PowerStream modelling stack is built for this new world.
If you’d like us to run a project-specific BESS revenue assessment, reach out.
Northstream Analytic 