In short. Gas substitution is not a fixed number. It rises with load and falls at light load, and it depends on engine, gas quality and controller limits. Estimate savings from the real load profile, and test step-load and fallback behaviour before relying on the set.
Substitution versus load
| Load band (illustrative) | Typical behaviour |
|---|---|
| Light (below ~30%) | Low or no gas; diesel carries combustion |
| Medium (~30–60%) | Rising substitution |
| High (above ~60%) | Highest substitution, within knock limits |
The maker’s data for your engine governs. Achievable diesel replacement in published material spans a wide range, so plan with conservative values.
Estimating savings
Weighted substitution = Σ (hours at load band × diesel use × substitution at that band) ÷ Σ (hours × diesel use)
Use a measured load histogram, not the nameplate rating.
Performance effects
- Efficiency and exhaust temperature differ from diesel-only.
- Possible derating at high ambient temperature or altitude. See ambient derating.
- Step-load response and frequency dip may change. See frequency dip.
Fallback
When gas pressure or quality falls, the controller returns to diesel. Test it under load, and check that the alternator and loads tolerate the transient.
Illustrative example
A plant spends 40 percent of hours below 30 percent load, where gas is barely used. A naive estimate of 70 percent replacement overstates savings; the weighted figure is much lower, and the payback is recalculated.
A worked savings example (illustrative)
A set consumes 120 litres of diesel per hour at average load. It runs 2,000 hours a year. Annual diesel use is 240,000 litres. Suppose a trial shows an average weighted substitution of 40 percent of diesel energy over the real load profile.
- Diesel avoided: 240,000 × 0.40 = 96,000 litres.
- Gas energy needed equals the energy in those litres, adjusted for any efficiency difference.
- Savings = avoided diesel cost − gas cost − added maintenance.
- Payback = conversion cost ÷ annual savings.
Substitute your own prices and measured fuel use. The point is the method, not the numbers.
Load behaviour in detail
| Behaviour | Explanation |
|---|---|
| Low substitution at light load | Lean mixtures burn poorly; controller limits gas |
| Higher substitution at medium and high load | Richer, hotter conditions support gas combustion |
| Ceiling at high load | Knock and temperature limits |
| Slower transient response | Gas in the manifold delays response to load change |
Managing load steps
- Sequence large motor starts.
- Allow the controller to raise the diesel share before a step.
- Use soft starters or drives on large motors.
- Test the worst-case step.
See motor starting and load shedding.
Derating
High ambient temperature and altitude can reduce allowable output, and gas operation may reduce margins further. Check the maker’s ratings for your site conditions. See ambient derating and altitude derating.
Fallback testing
Simulate loss of gas while the set is loaded. Record the frequency and voltage dip and the time to full diesel. Confirm the protected loads remain within tolerance.
The master engineering question
What does the real load histogram say about the achievable gas share?
What to send us
Send load logs and fuel records. Send your application details. See assessment.
Questions engineers ask
Is 90 percent replacement realistic?
Only at suitable load and gas; plan conservatively.
Does load step response change?
It can; test it.
What about low-load running?
Expect little gas use.
Do I need a load bank test?
Yes, to verify behaviour.
General guidance only. Values illustrative.