A flare pilot is a small flame with a very large responsibility. It sits at the tip of the stack, often no bigger than a candle, and its job is to light the relief stream the moment it arrives. When that flame goes out, or refuses to light at all, the flare is effectively blind: hydrocarbon goes to atmosphere unburnt, the site smells of it, and the safety case leans on a system that is not performing. The good news is that pilot failures are rarely mysterious. Almost every one traces back to wind, liquid, fuel gas quality, ignition hardware or a blocked tip.
A pilot must do two things: stay lit in whatever weather the site throws at it, and ignite the main flare stream reliably when a release occurs. Most designs use a shielded burner that mixes gas and air before combustion, with a flame retention shroud protecting the flame from the crosswind. Pilots are usually duplicated so one can fail without leaving the flare unguarded, and each one has a flame detector — a thermocouple, or a UV or IR sensor — reporting back to the control room.
That detection loop matters more than people expect. A lit pilot that reads as failed will trigger alarms and shutdowns, and it will send operators looking for a fault that is not there. So when you troubleshoot, separate the two questions: is the flame actually out, and is the instrumentation telling the truth?
Wind is the classic cause of pilot loss. A strong crosswind pushes the flame sideways out of the shroud, and if pilot gas flow is low, the flame simply cannot hold on. High winds around the stack can also create local low-pressure zones and turbulence that lift the flame off the burner. Rain and hail quench a small flame quickly, and snow or ice can block the air passages in the pilot shroud.
There is a subtler version of this problem. The flame may still be lit but held away from the thermocouple, so the panel reports a failure while a pilot is quietly burning a few centimetres to one side. If failures cluster on windy days, suspect aerodynamics rather than gas supply. Check the flame shield for damage, confirm the pilot is oriented as designed, and verify that pilot gas rate matches the commissioning data.
Even a small slug of condensate will put a pilot out. Fuel gas is not always as dry as the datasheet suggests, and liquids find their way in from wet gas, compressor lube oil carryover, glycol, or heavy ends condensing in cold pilot piping overnight. A knock-out drum that is not draining properly, or a filter that has never been changed, will do it too.
In colder climates, hydrate and ice formation in small-bore pilot lines is a real risk. Heat tracing, insulation and drip legs at low points are not luxuries. If your pilots fail most often on the first cold morning after a quiet spell, the problem is liquid, not ignition.
Pilots are fussy about what they burn. Gas with a high inert content has a lower heating value and a narrower band in which it will stay lit. Heavy hydrocarbons crack in the heat and deposit carbon on the pilot orifice, slowly strangling the flow until the flame starves. Wide swings in the fuel gas header — caused by regeneration cycles, compressor trips or other users taking gas — can drop pilot pressure below what the burner needs.
Pressure problems go both ways. Too little gas and the flame cannot be sustained; too much and it lifts off the burner or burns unstable. Undersized regulators, a pilot orifice that was changed during a modification, or a restricted pilot line will all produce the same symptom. Compare the pressure at the pilot panel with the pressure at the tip, and check both against the design figures.
If a pilot will not light at all, the ignition system is the first place to look. High-energy spark igniters and flame front generators both have known weak points: electrode gaps that erode or bridge with carbon, cracked ceramic insulators, moisture in the igniter housing, damaged high-tension cable, and poor earthing. A flame front generator also needs instrument air and gas at the right pressures, with a clean firing chamber, or the ignition shot never reaches the pilot.
Do not overlook the panel. A latched alarm, a timer out of step, a solenoid that has stopped stroking, or a switch left in the wrong position will defeat a perfectly healthy igniter. Watch the sequence from the control room and confirm each step actually happens.
Corrosion products, mill scale, sand and insect nests all find their way into pilot air passages and orifices. Carbon builds up inside the shroud until the flame can no longer breathe. Meanwhile, the detector that is supposed to confirm success may be misaligned, coated, or positioned where the flame never reaches it. Cleaning a pilot tip is often ten minutes of work that saves a week of head-scratching.
Most pilot reliability comes down to routine care. Keep the fuel gas clean and dry, maintain filters and drip legs on a schedule rather than after a failure, and record pilot gas pressure, orifice size and detector type so the next person is not guessing. After every turnaround, re-commission the pilots and confirm they light in the conditions they are meant to face.
Train operators to recognise what a healthy pilot looks like, because a steady flame and a steady signal are both part of the picture. And remember that flare systems are safety-critical. Work under a permit with the system isolated, purged and depressurised as required, and never change a pilot orifice, pressure setting or detector position without engineering review and proper management of change. If a pilot failure keeps returning after these checks, escalate it — that pattern usually points to a design or supply issue that needs more than field adjustment.
Photo: webandi / Pixabay