Insulation Heat Loss Calculator

How insulation heat loss is calculated

Heat leaving a hot insulated surface passes through two resistances in series: conduction through the insulation, then convection and radiation from the outer jacket to the surroundings. Total heat flow is the temperature difference divided by the sum of those resistances.

Cylindrical surfaces (pipes)

Per metre of pipe length, each insulation layer contributes:

and the outer surface contributes:

Flat surfaces (tank shells, duct walls)

Spheres and dished heads

Buried pipelines

The soil is treated as a semi-infinite medium containing an isothermal cylinder. The exact conduction shape factor gives, per metre of length:

where z is depth to the pipe centreline and D_o the outside diameter of the insulation. That resistance is added in series with the insulation. Ground-surface convection is neglected because soil resistance dominates at practical burial depths, and the relation is only valid while 2z > D_o. Use the undisturbed ground temperature at burial depth as the sink, not air temperature — in most climates it is several degrees cooler and far steadier than air.

The outer surface coefficient h_o

The surface coefficient is the sum of a convection and a radiation part, and both depend on the jacket temperature T_s — which is itself unknown. Radiation is handled with a linearised coefficient:

Convection is evaluated from the Rayleigh and Reynolds numbers using dry-air properties at the film temperature (T_s+T_a)/2, rather than from a fixed simplified formula:

• Vertical walls — Churchill–Chu, with the separate turbulent branch above Ra = 10⁹.
• Horizontal cylinders — Churchill–Chu, valid to Ra = 10¹².
• Spheres — Churchill, which correctly tends to Nu = 2 in the conduction limit.
• Horizontal plates — 0.54 Ra^1/4 and 0.15 Ra^1/3 for a heated face pointing up, 0.27 Ra^1/4 for a heated face pointing down. These swap over automatically on cold service, because a cold face pointing up produces the stable, suppressed boundary layer.
• Wind — Churchill–Bernstein for cylinders in cross-flow, Whitaker for spheres, flat-plate parallel flow for walls. Natural and forced contributions are blended as Nu³ = Nu_nat³ + Nu_forced³, so there is no discontinuity as the wind rises from zero.

Temperature-dependent conductivity

Insulation conductivity rises strongly with temperature, and each layer sits at a different mean temperature that is only known once the problem is solved. This calculator therefore updates k for every layer from that layer's own mean temperature inside the iteration. In a two-layer hot system the hot inner layer correctly runs at a much higher conductivity than the cooler outer layer. Assuming a single room-temperature k value is one of the most common and most costly errors in hand calculations — on a 400 °C line it can understate heat loss by well over a third.

How the surface temperature is found

Rather than repeatedly substituting until the numbers settle, the calculator solves the surface energy balance

by bisection between T_a and T_i. Because f is monotonic on that interval, convergence is guaranteed even for very hot surfaces, thick multi-layer systems, high wind or cold service — cases where simple successive substitution can oscillate or stall. The energy balance closes to better than 10⁻⁸ %.

Worked example you can check by hand

NPS 6 pipe (168.3 mm OD) carrying 150 °C process fluid, 50 mm mineral wool, painted finish (ε = 0.90), 35 °C still air:

Over a 30 m run at 8000 h/year that is roughly 1.7 kW insulated against 30 kW bare — about 13,800 kWh versus 244,000 kWh of delivered heat per year, before boiler efficiency.

Design checks the calculator applies

• Personnel burn risk — flags a jacket above 60 °C, the widely used industrial screening limit derived from ASTM C1055 contact-burn data for a metal finish.
• Condensation — on cold service, compares the jacket temperature with the ambient dew point from air temperature and relative humidity (Magnus relation). A jacket at or below dew point will sweat, and once moisture is inside the insulation its thermal performance is permanently degraded.
• Insulation service limit — flags any layer whose hot-face temperature exceeds the typical maximum service temperature of that material, or falls below its minimum.
• Critical radius — for a cylinder, insulation only reduces loss once the outer radius exceeds k/h_o, roughly 4–6 mm for common products in still air. On instrument tubing and small impulse lines a thin layer can increase heat loss, and the calculator says so.
• Burial validity — rejects a buried case where the depth to centreline is less than the insulation radius, instead of silently clamping it.
• Corrosion under insulation — screens the metal temperature against the susceptible bands in API 583 and AMPP/NACE SP0198, adjusted for service pattern and for whether the insulation touching the metal absorbs water.
• Economic thickness — sweeps stocked thicknesses and reports the one that minimises annualised capital plus the cost of heat lost, then raises it if a safety or condensation limit would be breached.

Insulation material properties used

Conductivity is modelled as k(T_m) = a + b·T_m + c·T_m² with T_m the layer mean temperature in °C. The values below are representative of the product family, fitted to typical published data. For detailed design, substitute the manufacturer's certified, tested conductivity curve — real products vary with density, binder and test method, and an ASTM C680 compliance calculation requires certified data.

Surface emissivity of finishes

Emissivity has a large effect on both heat loss and jacket temperature, and the two move in opposite directions. A bright aluminium jacket radiates poorly, so it loses less heat — but it therefore runs hotter to the touch than a painted jacket on the same insulation. In the worked example above, dropping ε from 0.90 to 0.09 cuts loss from 57.5 to 53.9 W/m but raises the jacket from 42.5 to 50.0 °C. That trade-off is exactly why burn-risk assessments must use the emissivity of the finish that will actually be installed.

Soil thermal conductivity for buried lines

A wet-season saturated case can more than double heat loss from the same buried line compared with a dry season, so a monsoon or high-water-table condition normally governs the design.

Assumptions, validation and limitations

Assumptions

• Steady state. No warm-up transients, cyclic operation or thermal mass effects.
• Pipe wall and internal film resistance neglected — the bare metal surface is taken to be at process temperature. For metal pipe and vessels carrying liquid or condensing steam this is worth well under 1 % of the total resistance. It is not valid for low-velocity gas service, plastic or lined pipe, or thick refractory walls, where the internal film can matter.
• Perfect installation. No gaps, compression, wet insulation, missing sections or through-metal supports. Real installed performance on an aged plant is commonly 10–30 % worse; a badly wetted or gapped system can be far worse than that.
• Thermal bridging is not modelled — pipe shoes, saddles, hangers, nozzles, manway necks, stiffener rings and ladder clips all conduct heat straight through the insulation.
• One-dimensional radial or normal heat flow in each geometry, with a uniform surroundings temperature equal to the air temperature. A jacket facing a hot furnace wall or a clear night sky radiates to a different effective temperature than that.
• Dry air at 1 atm for all surface properties.
• Ends, edges and corners are ignored on flat and spherical geometries.

Validation

The calculation engine is checked against closed-form and published results:

• Air properties reproduce standard tabulated values for conductivity, kinematic viscosity, Prandtl number and thermal diffusivity at 300 K and 400 K to within 2.5 %.
• Churchill–Chu vertical plate returns Nu = 147.1 at Ra = 1.813×10⁹, Pr = 0.690, matching the textbook value. Churchill–Chu horizontal cylinder returns Nu = 14.5 at Ra = 10⁶.
• With conductivity forced constant, the computed resistances reproduce the analytic forms t/k, ln(r₂/r₁)/(2πk) and (1/r₁−1/r₂)/(4πk) exactly, and series resistances add exactly across three layers.
• The linearised radiation coefficient converges to 4εσT³ as the temperature difference approaches zero, with no discontinuity at T_s = T_a.
• The buried-pipe soil resistance matches arccosh(2z/D)/(2πk) analytically.
• The surface energy balance closes to better than 10⁻⁸ % in every geometry, and results behave monotonically and in the right direction for thickness, wind, emissivity and reversed temperature gradient.

What this tool does not do

It is not an ASTM C680 compliance calculation, and it does not compute economic thickness, refractory linings, electrically traced or steam-traced lines, individually modelled fittings, jacket condensation on the underside of horizontal cold ducts, freeze protection time, or cool-down and warm-up transients. For heat-up time or process-side temperature drop along a line, pair this with a heat transfer and pressure drop calculation.

Corrosion under insulation (CUI)

Insulation saves energy and destroys steel. Every insulated system is an annulus that can collect water and then hold it against the metal, and CUI is one of the largest single contributors to unplanned loss of containment in process plant. The damage is invisible from outside, which is what makes it dangerous: the cladding looks fine right up to the point the pipe leaks.

This calculator screens each item against the temperature bands used in API 583 and AMPP/NACE SP0198. It is a screening step, not an inspection plan, and it deliberately reports a mechanism and a mitigation order rather than a single score.

Two different mechanisms, two different bands

Carbon and low-alloy steels corrode generally wherever water reaches warm metal. Austenitic stainless steels usually do not corrode generally at all — they crack, by external chloride stress corrosion cracking, which needs chlorides, moisture and tensile stress present at the same time. The two mechanisms peak in different places, so one temperature can be dangerous for one metallurgy and harmless for the other.

Different documents quote slightly different limits, and the figures above are the commonly applied ones rather than the only defensible set. Treat the band edges as soft.

Why the peak sits near the boiling point

Corrosion rate rises with temperature, but it needs liquid water. Below freezing there is no liquid; well above 120 °C water that gets in flashes off and the annulus runs dry. In between, water reaches the metal and stays there while the metal is hot, and that combination produces the highest rates. A 100 °C condensate line is therefore a far worse CUI prospect than a 400 °C steam line, which is the opposite of most people's intuition.

The corollary matters more than the rule. A line running above 175 °C is protected only while it is running. Shutdown, standby, steam-out and spared service all put the metal back into the band with water available, which is why hot lines are so often found corroded at a turnaround rather than during operation. Cycling makes it worse again: every heat-up drives vapour out of the annulus and every cool-down draws humid air back in, so a cycling system pumps far more water through itself than a steady one at the same temperature.

The mitigation order

The most common error is to treat insulation selection as the CUI control. It is not. API 583 treats the coating on the metal as the controlling barrier, with insulation and weather barrier as supporting measures. In order of effect:

1. Coat the metal. An immersion-grade epoxy or epoxy phenolic to roughly 200 °C, or thermal-spray aluminium for hotter or strongly cyclic service. The coating is what stands between water and steel once everything else has failed.
2. Keep water out. A continuous weather barrier, laps shingled so water sheds rather than wicks, and sealed terminations at every penetration, support, nozzle and flange break. Most CUI enters at a detail, not through the middle of a run.
3. Choose insulation that does not hold water. Closed-cell products such as cellular glass, and hydrophobic breathable blankets such as aerogel, do not act as a poultice. Open-cell and granular products hold water at the metal and release it slowly. Note the trade-off: closed cell does not absorb water, but water that does get in cannot drain either, so joint sealing matters more rather than less.
4. For austenitic stainless, control the chemistry. Specify insulation qualified to ASTM C795 and verify leachable chloride, fluoride, silicate and sodium to ASTM C871. C795 controls the ratio of ions rather than chloride alone, because silicate acts as an inhibitor.
5. Inspect where water collects, not at uniform intervals along a run: penetrations, supports and shoes, low points, dead legs, valve and flange boxes, dented or damaged cladding, and anywhere a previous repair broke the barrier.

On cold service the priority inverts. A line running below the ambient dew point condenses continuously, so it is permanently wet and the vapour barrier, not the insulation, becomes the corrosion control. The calculator raises the screening risk when it predicts a jacket at or below dew point for exactly this reason.

One limitation worth stating plainly: the screening takes the metal temperature to be the process temperature. That is correct for bare-wall metal piping in liquid or condensing service, but wrong for a refractory-lined or heavily fouled wall, where the metal can sit far from the process temperature and therefore in a different band.

Economic thickness of insulation

Insulation costs money once and saves money every hour. Add the annualised installed cost to the annual cost of the heat still being lost, and the total has a minimum. That thickness is the economic optimum. Thinner and you are buying fuel you did not need to buy; thicker and you are buying insulation that will not pay for itself.

The calculation

Total annual cost at thickness t:

where H is annual operating hours, η the efficiency of the heat source and p the price of delivered energy. The capital recovery factor converts a one-off installed cost into an equivalent annual charge:

for discount rate i over insulation life n. At 12 % over 10 years the factor is 0.1770, so each unit of installed cost carries an annual charge of about 18 %.

Installed cost is modelled in two parts, because they scale differently. Material cost follows the volume of insulation, which for a pipe grows as π(r_o² − r_i²) per metre. Cladding, banding and labour follow the finished outer area, 2πr_o per metre. Both rise with thickness, and the area term is why thick insulation gets expensive faster than a volume estimate alone suggests.

The answer is snapped to sizes you can buy

A calculated optimum of 47 mm is useless because nobody stocks it. The sweep runs over real thicknesses — ASTM C585 nominal sizes, or the metric sizes commonly stocked in India — so the recommendation is something a vendor can quote against. A 5 mm fine sweep is available if you want to see the true shape of the curve.

Three things the result will teach you

The curve is flat near the bottom. Total cost typically varies by only a percent or two across a wide band of thicknesses around the optimum. The calculator reports every thickness within 2 % of the minimum, and anything inside that band is economically equivalent. Choose within it on stock availability, cladding practicality and support design, not on the decimal place. Arguing about 90 against 100 mm is usually arguing about noise.

Economics is frequently not the binding constraint. On hot service the 60 °C touch limit often demands more thickness than money does; on cold service the dew point does. When that happens the calculator reports the unconstrained optimum, raises the recommendation to the cheapest thickness that actually passes, and names the constraint that bound it. A tool that reports economics alone will happily recommend a thickness that burns people.

At current energy prices the optimum is thicker than the old tables say. Legacy economic thickness tables were computed against fuel prices from a different era. Re-run the arithmetic at today's delivered energy cost and the answer usually comes out thicker than both the published table and what the plant actually installed. If the number looks surprisingly large, check the energy price and operating hours before assuming the calculation is wrong.

What is excluded, and why

• Uninsulated fittings. Their loss does not change with insulation thickness, so including them would add a constant to both sides and only obscure the optimum. Deal with them separately using removable jackets, which is nearly always a better investment than extra thickness on the straight run.
• Multi-layer installation cost. Above roughly 100 mm insulation is normally applied in two layers with staggered joints, which costs more to install than one layer of the same total thickness. If your quotation is per cubic metre of a single layer, the capital figures understate a thick recommendation.
• Maintenance, re-insulation and scaffolding over the life of the system, and any credit for reduced emissions or improved process control.

The default installed costs are indicative placeholders for Indian pipe work and must be replaced with your own quotations before the answer means anything. The physics in this tool is validated against textbook correlations; the prices are not, because prices are local.

Symbols and units

Related calculators

References

• ASTM C680 — Standard Practice for Estimate of the Heat Gain or Loss and the Surface Temperatures of Insulated Flat, Cylindrical, and Spherical Systems by Use of Computer Programs.
• ASTM C1055 — Standard Guide for Heated System Surface Conditions that Produce Contact Burn Injuries.
• ASTM C1057 — Practice for Determination of Skin Contact Temperature from Heated Surfaces Using a Mathematical Model and Thermesthesiometer.
• API 583 — Corrosion Under Insulation and Fireproofing. Source of the susceptible temperature bands and the mitigation hierarchy used in the CUI screening.
• AMPP/NACE SP0198 — Control of Corrosion Under Thermal Insulation and Fireproofing Materials: a Systems Approach.
• ASTM C795 — Standard Specification for Thermal Insulation for Use in Contact with Austenitic Stainless Steel.
• ASTM C871 — Test Methods for Chemical Analysis of Thermal Insulation Materials for Leachable Chloride, Fluoride, Silicate and Sodium Ions.
• ASTM C585 — Practice for Inner and Outer Diameters of Thermal Insulation for Nominal Sizes of Pipe and Tubing.
• ASTM C547, C533, C552, C591, C578, C534, C610, C1728 — material specifications for the insulation families listed above.
• Churchill, S. W. and Chu, H. H. S., correlating equations for laminar and turbulent free convection from a vertical plate and from a horizontal cylinder, International Journal of Heat and Mass Transfer, 1975.
• Churchill, S. W. and Bernstein, M., correlating equation for forced convection from gases and liquids to a circular cylinder in cross flow, Journal of Heat Transfer, 1977.
• Incropera, F. P. et al., Fundamentals of Heat and Mass Transfer — free and forced convection correlations, conduction shape factors.
• ISO 12241 — Thermal insulation for building equipment and industrial installations: calculation rules.
• BS 5422 — Method for specifying thermal insulating materials for pipes, tanks, vessels, ductwork and equipment.
• IS 14164 — Indian Standard, industrial application and finishing of thermal insulation materials: code of practice. Also a source of economic thickness guidance for Indian conditions.
• Economic thickness method: total-cost minimisation as set out in ISO 12241 and BS 5422, and in the Economic Thickness of Insulation approach implemented in NAIMA 3E Plus. Note there is no dedicated ASTM standard for economic thickness.