Low-Fire CO Spike on a Parallel-Positioning Burner
At 02:40, a 2,500 kg/h natural-gas firetube boiler at 800 kPa is modulating on header pressure to hold 700 kPa. The burner uses parallel positioning with a separate fuel-valve actuator and air-damper actuator, plus O2 trim in automatic. Logs show 4.6% O2 and 25 ppm CO at 65% firing. When the process drops and the burner steps to 20%, O2 falls to 1.1% and CO climbs to 650 ppm for about three minutes before recovering to 4.2% O2 and 30 ppm CO. The air actuator was replaced last year, but the damper linkage has visible play. Stack draft is steady at -0.07 kPa, flame signal is steady, and fuel pressure is constant. The operator proposes increasing the low-fire air position with the trim bias and continuing production. You are the only power engineer on shift.
1. What is the most probable mechanism for the transient O2 drop and CO spike, and why is it more likely than an O2 sensor fault or a fuel-pressure problem?
Read the worked answer
The key clue is timing: the problem occurs only during a downward load change and clears as the burner settles at 20%. At steady 65% firing, 4.6% O2 with 25 ppm CO is normal. At steady low fire, 4.2% O2 with 30 ppm CO is also acceptable. The dangerous condition is the three-minute transition, when O2 reaches 1.1% and CO reaches 650 ppm.
That pattern points to a temporary air deficiency caused by air-side linkage lag or hysteresis, not a steady air/fuel ratio error. On a parallel-positioning system, the fuel valve and air damper have separate actuators. If the air actuator or damper linkage sticks, has backlash, or strokes more slowly than the fuel valve on a decreasing load, the fuel valve closes toward the 20% position while the damper briefly remains too far open or too far closed. Here the O2 falls, so the air is lagging on the low side: the damper is not reaching its commanded low-fire position quickly enough, or the linkage is hanging and then slipping. The visible play supports this.
An O2 sensor fault is less likely because the event is repeatable with firing-rate changes, and the reading recovers to a plausible value. A sensor would not normally produce a CO spike that tracks the load change. Fuel-pressure instability is also less likely because fuel pressure is constant and high-fire combustion is normal. The CO spike is the real concern: it shows incomplete combustion and unburned fuel, which is both an efficiency loss and a potential explosion hazard if unburned gas accumulates.
2. Describe a safe diagnostic sequence that separates linkage/actuator hysteresis from O2 analyzer error and from fuel-valve stroke problems, without opening the burner while firing.
Read the worked answer
Place the O2 trim in manual at a known bias. Do not change the trim to mask the problem. Record the boiler load, fuel pressure, draft, flame signal, O2, and CO at steady 65%, 40%, and 20% firing.
Verify the analyzer. Check the sample probe location, sample line for cracks or water traps, and compare the permanent analyzer against a calibrated portable analyzer at the same point. A leaking sample line can dilute the sample and raise O2, so confirm the reading before blaming the linkage.
Perform a step test. With the boiler at steady 65%, command a step to 20% and record O2, CO, and the commanded versus actual fuel-valve and air-damper positions every 5 to 10 seconds. Repeat from 20% to 65%. Use an inclinometer, a marked linkage, or the actuator feedback if available. Watch the damper and linkage while a second person observes the analyzer.
Check mechanical condition. Look for loose ball joints, worn pins, a binding damper shaft, a damper blade hitting its stop, and actuator stroke time. Compare the air actuator stroke time with the fuel-valve actuator. A damper that is slow on closing will show up as air lag on a decreasing load.
Check the fuel valve separately. Confirm it strokes smoothly and reaches its low-fire position. If the air damper reaches position but fuel remains high, the fuel valve is the problem; if the air damper lags, repair the linkage and have a qualified burner technician re-characterize the low-fire position. After any repair, repeat the step test and perform a combustion test at low, mid, and high fire.
3. What interim operating limits would you set if production must continue, and what would cause you to take the boiler off line?
Read the worked answer
Interim operation should be treated as a controlled, documented exception, not a return to normal. I would keep the O2 trim in manual and prevent rapid load changes by fixing the boiler at a steady load or limiting modulation. I would raise the low-fire limit to about 30-35% until the linkage is repaired, because the 20% position is where the air lag produces the CO spike. I would increase excess air by about 1-2% O2 across the operating range, but only after verifying with a portable analyzer that CO actually falls at every firing rate.
I would set alarms and a clear trip point: CO above 200 ppm for more than one minute, O2 below 2.5% at steady load, flame signal instability, or repeated CO above 400 ppm during load changes. I would log O2 and CO at least every hour and after every load change. The boiler should not be left unattended, and the chief engineer should authorize the interim plan in writing.
I would take the boiler off line immediately if the CO spike becomes sustained rather than transient, if O2 falls below 1.5% at steady load, if the flame scanner signal becomes erratic, if there is any sign of afterburning or puffing, or if the linkage cannot be made to repeat reliably. A 650 ppm CO transient is already a warning; a second event of the same size without correction would be my shutdown trigger. The proper repair is mechanical linkage correction and burner re-characterization, followed by a repeat step test.