A combined cycle unit's heat rate does not move in a straight line as load changes, so the cost of the next megawatt at 60 percent load can be meaningfully different from the cost at 95 percent load, and a dispatch decision that ignores that curve leaves real revenue sitting on the table every hour of the day. Operations directors are expected to balance that nonlinear heat rate curve against fuel price swings, emissions allowance costs, and capacity market obligations, often inside the same five-minute window the grid operator is watching. Get that balance wrong and the plant either burns more fuel than the megawatt was worth, or under-generates against a capacity commitment it is contractually required to meet. A static heat rate table built once a year and a fuel price assumption from last quarter cannot keep pace with a market that reprices fuel, emissions, and capacity value continuously. iFactory's AI-driven platform models each unit's real heat rate curve against live market signals, so every dispatch decision reflects what the megawatt is actually worth right now. Book a demo to see live dispatch economics modeled against your fleet's own heat rate curves.
AI-Driven · Combined Cycle Dispatch · Market Economics
Every Dispatch Decision Is a Bet on What the Next Megawatt Is Actually Worth
iFactory models your unit's real heat rate curve against live fuel cost, emissions allowance pricing, and capacity obligations, so dispatch decisions maximize revenue per MWh instead of following a static heat rate table.
Nonlinear
Heat rate curves rise and fall unevenly across the load range, not on a straight line
5-Minute
Typical dispatch interval where fuel, emissions, and capacity value can all reprice
4 Inputs
Fuel cost, heat rate, emissions allowance price, and capacity obligation all move independently
Real-Time
Market signals the plant is dispatched against, versus the static assumptions dispatch tables use
The Dispatch Problem
Four Variables That Move Independently, One Decision Made Every Interval
A dispatch call that looks right against yesterday's fuel price can be the wrong call an hour later, because none of the four variables driving the economics move on the same schedule.
Curve Shape
The Heat Rate Curve Bends, It Doesn't Slope
Efficiency typically improves through the mid-load range and can fall off again near the top end, so the true cost of the next megawatt depends on exactly where the unit is sitting on its own curve, not a single average heat rate.
Fuel Cost
Fuel Price Moves Faster Than the Dispatch Table Updates
A dispatch table built from last month's fuel index can misprice generation the moment natural gas prices move, and manual updates rarely happen fast enough to keep the plant dispatched at the correct margin.
Emissions Cost
Emissions Allowance Pricing Changes the Real Margin
The cost of the emissions allowances tied to each megawatt is a real, variable cost of generation, and ignoring its current price can make a dispatch decision look profitable on paper when it is actually running at a loss.
Capacity Value
Capacity Market Obligations Set a Floor, Not a Suggestion
A unit with a capacity market commitment has to generate through certain conditions regardless of that hour's energy margin, and dispatch models that ignore this obligation can recommend a curtailment that triggers a non-performance penalty.
Cost Breakdown
What Actually Makes Up the Cost of Your Next Megawatt
Four cost components stack up behind every dispatch decision, and the mix between them shifts hour to hour as fuel and emissions markets move.
Startup and Ramp Cost (Amortized)
Illustrative cost mix for a mid-merit combined cycle unit. The actual split shifts continuously as fuel and allowance prices move, which is exactly why a fixed dispatch table falls out of date.
How the Platform Works
From Live Market Signal to Dispatch Recommendation in Four Steps
1
Model the Real Heat Rate Curve
The platform builds each unit's actual heat rate curve from operating data across its full load range, replacing the single average number most dispatch tables rely on.
2
Ingest Live Market Signals
Fuel price, emissions allowance value, and grid price feed into the model continuously, so the cost picture updates as fast as the market does.
3
Apply Capacity Obligations as Constraints
Capacity market commitments are built into the model as hard constraints, so no recommendation ever suggests a curtailment that would trigger a non-performance penalty.
4
Recommend the Load Point
The platform recommends the load point that maximizes revenue per MWh at that moment, updated continuously as inputs change, instead of once per dispatch shift.
Decision Reference
How Dispatch Logic Changes as Market Conditions Shift
The right dispatch call depends on which combination of fuel price, demand, and obligation status the plant is facing at that moment.
| Market Condition |
Fuel Price |
Grid Demand |
Typical Optimal Response |
| Peak Demand Window |
Elevated |
High |
Run near best-efficiency load point, capture the energy margin while it's available |
| Off-Peak, No Obligation |
Elevated |
Low |
Reduce load toward minimum stable point, avoid running at negative margin |
| Off-Peak, Capacity Obligation Active |
Elevated |
Low |
Hold the contracted floor even at thin margin, avoid a non-performance penalty |
| Low Fuel Price, Moderate Demand |
Low |
Moderate |
Push toward higher load where the curve still favors efficiency, capture wider margin |
Dispatch Readiness
The Six-Point Dispatch Economics Scorecard
1
Each unit's heat rate curve is modeled from real operating data across its full load range, not a single average figure
2
Fuel cost inputs update continuously against a live index instead of a fixed monthly or quarterly assumption
3
Emissions allowance pricing is included as a real variable cost in every dispatch recommendation
4
Capacity market obligations are enforced as hard constraints, never overridden by a short-term margin calculation
5
Startup and ramp costs are amortized into the recommendation instead of ignored as a sunk cost
6
Revenue per MWh generated is tracked against the recommendation, so the model's accuracy is measured, not assumed
Stop Dispatching Off a Table That Was Right Last Quarter.
Live heat rate modeling, market signal ingestion, and capacity obligation constraints, configured against your fleet's own units and contracts.
From the Field
What Changes When Dispatch Follows the Market Instead of a Table
We were dispatching two combined cycle units off a heat rate table that got refreshed once a quarter, and honestly most weeks nobody had time to check whether the fuel price assumption behind it still held. When gas prices moved fast during a cold snap, we found out after the fact that we had been running one unit closer to its capacity floor than the margin actually justified, while the other unit had room to pick up more load at better efficiency and nobody caught it in time. Once we connected live fuel and allowance pricing against each unit's actual heat rate curve, the recommended load points started shifting within the same dispatch interval the market moved, not a quarter later. The capacity obligation constraint gave our dispatch team confidence that the system would never recommend cutting below our contracted floor, which was the concern that had kept us on the manual table for as long as we had been.
— Operations Director, Independent Power Producer, Two-Unit Combined Cycle Fleet
Conclusion
The Market Reprices Every Interval. Your Dispatch Table Shouldn't Lag Behind It.
Fuel cost, emissions allowance value, and capacity obligations move on their own schedules, and a dispatch decision built on a table that only gets refreshed occasionally is, by definition, working from stale numbers most of the time. The nonlinear shape of the heat rate curve makes this worse, because the true cost of the next megawatt depends on exactly where the unit sits on that curve at that moment.
iFactory's AI-driven platform keeps the heat rate model, the market signals, and the capacity constraints connected continuously, so every dispatch recommendation reflects what the megawatt is actually worth right now. Book a demo to see it modeled against your own fleet's dispatch economics.
Frequently Asked Questions
Combined Cycle Dispatch Optimization — What Operations Directors Ask
How is a heat rate curve different from the single heat rate number in the OEM datasheet?
The OEM datasheet number is typically a single design-point figure, usually at or near full load, under specific ambient conditions. In actual operation, efficiency changes continuously as load, ambient temperature, and equipment condition shift, which means the real cost of generating the next megawatt at 70 percent load can be noticeably different from the datasheet number. Modeling the full curve from real operating data gives dispatch decisions a far more accurate cost basis than a single design-point figure ever could.
Book a demo to see your unit's actual heat rate curve modeled.
Does the platform override our capacity market commitments to chase a better energy margin?
No. Capacity market obligations are built into the model as hard constraints, not preferences that can be traded away for a better hourly margin. The platform will never recommend a load point that would drop the unit below its contracted capacity floor, even during hours where the pure energy margin on paper looks unattractive, because the cost of a non-performance penalty is factored into the decision.
How often does the dispatch recommendation update against market changes?
Market signals including fuel price and emissions allowance pricing feed into the model continuously, so the recommended load point can shift within the same dispatch interval the underlying market moves, rather than waiting for a scheduled table refresh. This is the core difference from a manually maintained dispatch table, which is typically only as current as its last update cycle.
Can this be used across a mixed fleet with different unit types and different capacity contracts?
Yes. Each unit's heat rate curve, contract terms, and capacity obligations are modeled individually, so a fleet with different combined cycle configurations and different market commitments can still be dispatched from one connected view rather than a separate spreadsheet per unit.
Contact support to discuss connecting a mixed fleet.
What data does a plant need before dispatch optimization can start?
Most of what's needed already exists in the plant's historian and control system, including load, fuel flow, and output data used to build the heat rate curve, along with contract terms for capacity obligations. The platform connects to those existing sources and live market feeds rather than requiring new instrumentation, so most plants can begin generating dispatch recommendations well before a lengthy integration project would otherwise be complete.
Know What the Next Megawatt Is Actually Worth Before You Dispatch It.
Live heat rate modeling, real-time fuel and emissions pricing, and capacity obligation constraints, configured for your fleet's own units and contracts.
