Flare Radiation and Siting — API 521 Thermal Contour Analysis

Introduction

A flare header sizing exercise tells you the diameter of the pipe carrying relief gas to disposal. It does not tell you how tall the stack needs to be, where to put it, or how far the manned exclusion zone reaches. Those answers come from a separate calculation — thermal radiation analysis — and they are what shape the plot plan.

API Standard 521 (Pressure-relieving and Depressuring Systems) gives the framework: a set of permissible radiation levels for different exposure scenarios, and several validated calculation methods for predicting the radiant flux at any point around the flare. This post works through the analysis from first principles, walks the API 521 Table 5 limits, and shows how the 1500 BTU/hr·ft² contour drives stack height and platform layout. The stack is the last link in the relief chain covered in the pressure relief and flare systems overview.

The Radiation Limits — API 521 Table 5

The starting point is the table of permissible radiation intensities, which has barely changed since the standard's first edition:

The thresholds reflect human-tolerance research from the 1950s–60s — at 4.73 kW/m², an unshielded person experiences pain within seconds and second-degree burns within around half a minute; the graded durations exist because clothing buys time but does not change the physics. At 1.58 kW/m², an unshielded person in normal coveralls can stand the heat indefinitely. The shorter-duration tiers assume the worker is performing a defined emergency action or moving away from the source — structural and equipment heat tolerance is a separate assessment against far higher fluxes, and does not belong in this table.

For most facility designs, the governing contour is the 1500 BTU/hr·ft² (4.73 kW/m²) limit at any location personnel may need to occupy during a relief event. This is what determines the radius around the flare that must be kept clear of routine occupancy.

What Drives the Radiation

The radiant flux at a receiver point is the product of four factors:

q = (τ × Q_rad) / (4π × D²)

Where:

• Q_rad = heat liberated as radiation (kW) — the radiant fraction F_rad applied to the total heat release
• τ = atmospheric transmissivity (typically 0.6–0.85 at typical humidity, 1.0 in dry air at short range)
• D = distance from the flame centre to the receiver

An isotropic point source has no geometric view factor — if you find yourself multiplying by both a radiant fraction and a view factor in this equation, one of them is being double-counted.

The radiation source term Q_rad depends on the flame heat release (mass flow × heating value) and the radiant fraction F_rad — the share of the combustion heat radiated as thermal energy rather than carried away in the combustion gases. F_rad varies with fuel composition and flame geometry:

Heavier fuels burn smokier and hotter — the soot in the flame is the dominant radiator. This is why a propane-rich relief produces a hotter exclusion zone than the same mass flow of methane.

The Geometry — Flame Length, Tilt, and Centre

The flame is not a point. API 521 gives correlations for flame length as a function of jet exit velocity (in the flare design clause — §5.7 in recent editions):

• For a sonic exit (typical at high relief mass flow): flame length L ≈ 0.00326 × Q^0.478 (L in metres, Q in W)
• Subsonic exit, the length scales with the Froude number and exit diameter

For a substantial relief event, flame lengths of 30–60 metres are typical from an offshore flare boom; 60–120 metres from a large refinery elevated flare. The flame tilt under wind is significant — at 10 m/s crosswind, a flare flame can lean over 30–45° from vertical.

For the radiation model, the convention is to treat the flame as a point source at the flame centre, located at the geometric centre of the visible flame at the design wind condition. The centre lies above the tip and downwind of it by the tilt distance — for a 60 m flame at 30° tilt, the centre is roughly 26 m above the tip and 15 m downwind of it.

The conservative practice is to evaluate radiation at the least-favourable wind direction — i.e. the direction that puts the flame centre closest to the receiver point you care about. For an offshore platform with the flare on a boom, this is usually the crosswind condition that brings the flame back toward the platform.

The Three Common Calculation Models

Simple point-source (Hajek-Ludwig)

The conservative starting point. Treat the flame as a single point at the flame centre, all radiation emitted isotropically, attenuated by atmospheric transmissivity. Easy to spreadsheet, conservative, and the basis for screening.

Brzustowski-Sommer

The most-cited refinement — still a point-source method, but a smarter one, with corrections for:

• Flame length and tilt as a function of jet exit velocity and wind speed
• The source located at the flame epicentre — found from the jet trajectory to the lower-flammability envelope under wind tilt — rather than at the visible-flame midpoint
• Atmospheric transmissivity as a function of humidity and path length

(Genuine solid-flame treatments, which model radiation from the flame surface — e.g. Chamberlain — are a different model family, used mainly for detailed dispersion-and-radiation studies.)

Brzustowski-Sommer typically gives results 15–30% less conservative than the simple point-source model — useful when the screening result is close to the limit and you do not want to over-build the stack.

Detailed CFD

A full computational fluid dynamics simulation of the combustion, plume, and radiative transfer. Used for unusual geometries (e.g. flares on enclosed platforms, ground flares, very high-pressure releases) or where the screening result is genuinely uncertain. Resource-intensive and only justified when the simpler methods reach an unworkable answer.

The Contour That Sets the Plot Plan

For an offshore platform, the typical workflow:

1. Establish the relief load at the governing contingency (general power failure or single-zone fire case, per the flare network analysis).
2. Compute the heat release: mass flow × LHV.
3. Estimate flame length at design exit velocity per API 521.
4. Pick the wind condition that produces the worst case for each receiver point (helideck, drill floor, accommodation, escape route).
5. Compute the radiation at each point using Brzustowski-Sommer (or simple if screening generously).
6. Compare to the limit: 1500 BTU/hr·ft² at any manned area during the design relief event; 500 BTU/hr·ft² for routinely occupied areas during normal pilot operation.
7. If exceeded, increase stack height (or boom length) until the limit is satisfied. Re-iterate the flame centre location at the new height.

The 1500 BTU/hr·ft² contour at the design relief is the contour that must not touch the helideck, the accommodation block, the lifeboats, the muster point, or any escape route. The 500 BTU/hr·ft² contour at the pilot flame (a much smaller, continuous radiation source) must not touch routinely occupied areas at all.

These two contours, mapped onto the plot plan, are what drive stack height. Typical offshore boom lengths and tip heights are 40–90 m, set by exactly this calculation.

The Smokeless Tip Question

Steam-assisted and air-assisted flare tips reduce visible smoke by entraining additional air into the combustion zone. The radiation consequence is significant:

• Smokeless (steam or air) tip: F_rad reduced typically by 30–40% versus open-pipe. Lower exclusion radius for the same gas flow.
• Open-pipe tip: highest F_rad, simplest mechanically, lowest maintenance.

For offshore platforms where plot space is at a premium, a smokeless tip can shave several metres off the required stack height. For onshore plants with more space, the choice often falls to operating cost (steam consumption) and reliability (smokeless tips have more parts).

The Sterilised Zone

Beyond the radiation contour, API 521 also implicitly defines a sterilised zone at grade — the area where:

• Vegetation must be cleared — sustained radiation desiccates it, and piloted ignition of dry cellulosic material begins at roughly 12 kW/m², a flux the near-field of the flare base can reach during low-flow events.
• Buried services and shallow utilities must be protected from ground heating.
• Surface roads, pipe racks, and structural steel must withstand the design radiation without yielding.

For onshore terminals, the sterilised zone often extends 30–80 metres around the flare base. Operators clear vegetation, install heat-resistant paint on nearby steel, and route services accordingly.

Worked Example — Offshore Platform Flare Boom

Scenario: A platform with a 60-metre flare boom carrying a 50 t/h relief at design contingency. Gas composition: mainly C1-C3, LHV ≈ 50 MJ/kg, F_rad ≈ 0.22. Wind 10 m/s, transmissivity 0.7.

Heat release: Q = 50,000 kg/h × 50,000 kJ/kg / 3600 = 694 MW total, of which Q_rad = 0.22 × 694 = 153 MW radiated.

Flame length at sonic exit, approximate: with Q = 694 MW = 694 × 10⁶ W, L ≈ 0.00326 × (694 × 10⁶)^0.478 ≈ about 55 m for a steam-assisted tip on this scale.

Flame tilt at 10 m/s wind: roughly 35° from vertical. Flame centre ≈ 22 m above tip (27.5 m half-length × cos 35°), ≈ 16 m downwind (× sin 35°).

Receiver point: helideck, 50 m horizontally inboard of the boom root, 25 m above sea level. Boom tip is at +95 m above sea level, 60 m outboard of the boom root; the flame centre sits at +117 m, 76 m outboard. Horizontal separation to the helideck is therefore 76 + 50 = 126 m, vertical 117 − 25 = 92 m, and the line-of-sight distance D = √(126² + 92²) ≈ 156 m.

Point-source radiation at the helideck:

q = 0.7 × 153,000 / (4π × 156²) ≈ 0.35 kW/m² ≈ 110 BTU/hr·ft²

This is far below the 1500 BTU/hr·ft² emergency-action limit, and comfortably below even the 500 BTU/hr·ft² continuous-occupancy threshold. The 60-metre boom is conservative for this relief load.

If the boom were shorter — say 40 m — the flame centre moves ~20 m closer: D ≈ 140 m and the design-relief flux at the helideck rises only to about 0.43 kW/m² ≈ 140 BTU/hr·ft² — still comfortable. The binding checks for a shorter boom are usually elsewhere: work areas near the boom root at design relief, and the continuous pilot-flame contour over routinely occupied areas — each evaluated against its own limit and its own source term. The trade is real, but it is those receivers, not the helideck, that decide it.

Common Pitfalls

• Computing radiation at the design wind speed only. The least-favourable wind direction is often the one that puts the flame closest to the receiver — not necessarily the design speed.
• Forgetting the pilot flame radiation. The pilot runs continuously; the 500 BTU/hr·ft² continuous-occupancy limit applies to it, not to the design relief.
• Using point-source for very low stacks. When the flame length is comparable to the receiver distance, point-source over-predicts by 50–100%. Use Brzustowski-Sommer.
• Ignoring flame tilt. A wind-tilted flame can bring the radiation hotspot inside the platform perimeter even though the tip is well outside it.
• Treating F_rad as fuel-independent. A propane-rich relief radiates twice as much per kg as a methane relief — the exclusion zones are not interchangeable.
• Not checking the helideck approach path. Helicopters approach into wind, which is often the same direction the flame is leaning — the approach path can pass through a contour that the static helideck does not.

Conclusion

Flare radiation analysis is what turns a relief mass flow into a piece of geometry. The flare network sizes the header; the API 521 thermal calculation sizes the stack, sets the boom length, and draws the contour that determines where the helideck, the accommodation block, and the escape routes can go.

A radiation analysis done well produces a stack tall enough to put the 1500 BTU/hr·ft² contour clear of every manned area at the design relief, and a 500 BTU/hr·ft² pilot-flame contour clear of routinely occupied areas. Done poorly, it produces a platform where the muster point is in the radiation envelope of the very flare the crew is supposed to be moving away from.

The discipline is simple — pick the right wind condition, the right flame geometry, the right radiant fraction, and the right receiver points. The arithmetic follows.