Permanently installed machines burning highly variable raw gas that a factory calibration can only survive, never exploit. This is the strongest case for site-specific combustion calibration — and our primary market.
Pipeline-quality natural gas is a managed product: heating value, Wobbe Index and contaminant limits are held inside tariff specifications. Raw produced gas is whatever the formation delivers — methane diluted or enriched by ethane, propane and C₄+ condensable fractions, ballasted with CO₂ and N₂, and frequently sour with H₂S. Two wells in the same field can supply gases with substantially different heating values, stoichiometric air requirements, flame speeds and adiabatic flame temperatures. Fuel-flexible generation equipment is specified to accept heating values spanning roughly 350 to 2,500 BTU/scf — a 7:1 volumetric energy range — with tolerance for very high CO₂ fractions and significant sour-gas concentrations.
That acceptance window is a survival specification, not an optimization. A control schedule that must remain safe everywhere inside a 7:1 fuel envelope is, by construction, conservative at every single point inside it.
For reciprocating gas engines, the governing fuel-quality parameter is the methane number (MN) — the gaseous-fuel analogue of octane, with pure methane at 100 and hydrogen at 0. Heavier hydrocarbons degrade it rapidly: rich associated gas with meaningful propane and butane content can fall well below the MN 70–80 window that high-BMEP, high-compression stationary engines are designed around.
When the delivered MN falls below the engine's design point, the options are all bad: end-gas knock, ignition-timing retard (with the fuel-consumption penalty that follows), or an outright power derate — in severe cases to a fraction of nameplate. Because installations must be adjusted for the least knock-resistant gas they may ever see, most run permanently detuned for a worst case that arrives rarely, if ever. A site-specific calibration built on measured composition, with knock margin actively managed rather than statically reserved, recovers that headroom.
On any fixed-orifice or venturi fuel system, the Wobbe Index directly sets thermal input — which means fuel-quality drift silently re-tunes the engine. A Wobbe excursion from 48 to 54 MJ/m³ on a lean-burn engine calibrated at λ = 2.0 drives the mixture to roughly λ = 1.78: hotter combustion, a NOx excursion, and a simultaneous collapse in methane number that can push a full-load engine into heavy knock. The opposite swing leans the mixture toward λ ≈ 2.3 — misfire territory, with ignitable mixture passing into the exhaust. Wellhead gas does this routinely. The correct response is composition-aware feed-forward control: measure or infer the gas, reschedule fuel, air, and ignition before the excursion, not after the trip.
Small industrial gas turbines built for wellhead service carry the same compromise in different clothing. The combustor and fuel system are engineered for enormous fuel flexibility, but firing temperature schedules, surge margins and emissions margins are set to protect the fleet across the entire fuel and ambient envelope. A permanently installed unit on a characterized gas stream doesn't need fleet-wide margins. Site calibration optimizes fuel mass flow, combustor equivalence ratio and firing temperature against the measured gas — and releases generator load when compressor margin, turbine temperature and emissions demonstrate it is safely available.
Gas-turbine and engine output ride on inlet air density: temperature, absolute barometric pressure and humidity. Ratings are quoted at ISO conditions — 15 °C, sea level, 60% RH — that a Canadian Prairie site sees a handful of days a year. Winter air at −30 °C is dramatically denser than a +30 °C summer afternoon; on a 300-kW-class turbine the swing in available power is tens of kilowatts. Site calibration cannot change air density, but it prevents the unnecessarily conservative derating that generic ambient-correction curves impose, exploiting cold dense air for additional load and concentrating on best achievable heat rate when summer density limits output.
At wellheads where associated gas would otherwise be flared, the fuel is nearly free — which changes the objective function, not the physics. Every point of efficiency converts more of a finite, declining gas stream into sellable or usable electricity instead of radiated heat, reduces flaring volumes, and extends the economic generation life of the well as production declines. Efficiency here isn't a fuel-cost line item; it is production capacity.
On pipeline-quality gas the recoverable margin is small — a few tenths of a point. On typical raw well gas, an engineering estimate of 0.5–1.5 absolute efficiency points (1.5–5% less fuel energy per kWh) is defensible. On highly variable, diluted or sour streams currently forcing conservative operation, 1–3 points may be recoverable — with additional benefit frequently appearing as restored available output rather than efficiency. Actual numbers are established by controlled baseline-vs-optimized testing on your machine and your gas.
| Operating condition | Generic / broad calibration | Site-optimized calibration |
|---|---|---|
| Pipeline-quality natural gas | ~32–33% ηₑ | ~32.5–33.3% ηₑ |
| Moderately variable raw well gas | ~31–32.5% ηₑ | ~32–33% ηₑ |
| Highly variable / diluted well gas | ~29–32% ηₑ | ~31–33% ηₑ |
| Extreme acceptable fuel | ~27–31% or derated | ~30–32% ηₑ |
Engineering estimates for a 300-kW-class recuperated gas turbine; established per-site by controlled testing, not assumed.
MNKnock resistance of a gaseous fuel (methane = 100, hydrogen = 0). High-BMEP stationary engines typically require MN ≥ 80; heavier hydrocarbons in associated gas pull it down fast. Below design MN: knock, timing retard, or derate.
WI = HHV/√SGThermal input through a fixed orifice — the fuel-interchangeability parameter. WI drift on a fixed-orifice system re-tunes λ without anyone touching the engine. Identical WI does not guarantee identical combustion behaviour.
λ = AFR/AFRstoichNormalized air/fuel ratio. Lean-burn gas engines run λ ≈ 1.7–2.0, boxed between the misfire limit (too lean) and the knock/NOx limit (too rich). Fuel-quality drift moves λ; calibration decides how the controller moves it back.
Deliberate output reduction to protect the machine from knock, overtemperature or instability on off-spec fuel. Often applied fleet-wide and statically — meaning most sites permanently give up output for a worst case they rarely see.
Gas produced alongside oil. Composition-rich (low MN), frequently sour, historically flared. Converting it to power makes efficiency a production parameter, not a cost parameter.
15 °C, 101.325 kPa, 60% RH — the reference point for nameplate output. Your site is not ISO. Cold dense air adds recoverable output; hot thin air demands heat-rate optimization at reduced load.
One extended gas analysis and the machine model is enough for a first-pass engineering estimate of what your site is leaving on the table.
Start the Conversation