LMTD Simulator — Log Mean Temperature Difference for Counter and Parallel Flow

Parameters

Hot inlet T_h_in

°C

Hot outlet T_h_out

°C

Cold inlet T_c_in

°C

Cold outlet T_c_out

°C

Defaults (T_h_in=150°C, T_h_out=90°C, T_c_in=25°C, T_c_out=80°C) yield counter-flow LMTD ≈ 67.5 K, parallel-flow LMTD ≈ 45.5 K and a counter / parallel ratio ≈ 1.48. Physically inconsistent combinations (T_h_in > T_h_out > T_c_in, T_c_out > T_c_in, T_h_out > T_c_out in parallel flow violated) are flagged with "—" or "invalid".

Results

—

Counter-flow LMTD

—

Parallel-flow LMTD

—

Counter / Parallel ratio

—

Counter-flow advantage

Temperature profiles (counter / parallel)

Top half = counter flow: hot stream left-to-right, cold stream right-to-left / Bottom half = parallel flow: both streams left-to-right / Red = hot stream / Blue = cold stream / Green dashed = ΔT_1, ΔT_2 at the two ends

LMTD/ΔT_1 vs ΔT_2/ΔT_1

Horizontal: end-difference ratio ΔT_2/ΔT_1 (0.05 to 1.0) / Vertical: LMTD/ΔT_1 (dimensionless) / Blue: LMTD function (1−r)/ln(1/r) / Yellow dot: current counter-flow / Orange dot: current parallel-flow / Counter flow usually sits at higher r (close to 1) and a larger LMTD

Theory & Key Formulas

The Log Mean Temperature Difference is the representative temperature difference along an exchanger whose local ΔT varies with position, expressed in terms of the end values $\Delta T_1, \Delta T_2$:

$$\mathrm{LMTD} = \frac{\Delta T_1 - \Delta T_2}{\ln(\Delta T_1/\Delta T_2)}$$

Counter and parallel configurations pair the end differences differently:

$$\Delta T_1^{\text{co}} = T_{h,\text{in}} - T_{c,\text{out}},\quad \Delta T_2^{\text{co}} = T_{h,\text{out}} - T_{c,\text{in}}$$ $$\Delta T_1^{\text{pa}} = T_{h,\text{in}} - T_{c,\text{in}},\quad \Delta T_2^{\text{pa}} = T_{h,\text{out}} - T_{c,\text{out}}$$

The total heat transfer rate, integrated along the area, is driven by LMTD:

$$Q = U\,A\,\mathrm{LMTD}$$

$U$ is the overall heat-transfer coefficient (W/m²K), $A$ the heat-transfer area (m²). At the removable singularity $\Delta T_1 = \Delta T_2$ the limit value $\mathrm{LMTD} = \Delta T_1$ is used.

What is the LMTD Simulator?

🙋

Whenever I size a heat exchanger the LMTD keeps showing up. What does it actually represent, and how is it different from a plain average temperature difference?

🎓

Good question. Inside an exchanger the local ΔT changes from inlet to outlet — large at one end, small at the other. The LMTD is not the arithmetic average, but a logarithmic one: LMTD = (ΔT_1 - ΔT_2) / ln(ΔT_1/ΔT_2). It is what you get by integrating dq = U dA ΔT(x) along the heat-transfer area; the log emerges naturally from exponential decay of the local ΔT. With the defaults of this tool (T_h_in=150C, T_h_out=90C, T_c_in=25C, T_c_out=80C) you should read counter LMTD about 67.5 K and parallel LMTD about 45.5 K.

🙋

A 22 K difference for the same inlets and outlets is huge. Both layouts move the same amount of heat, right? Why do they need such different LMTDs?

🎓

That is the core of it. In counter flow ΔT stays comfortably large at both ends: 150−80 = 70 K at one end and 90−25 = 65 K at the other. In parallel flow ΔT is huge at the inlet (150−25 = 125 K) but collapses to 90−80 = 10 K at the outlet, so the integrated average falls. Look at the top half of the temperature-profile chart: the counter-flow lines stay nearly parallel as they descend, whereas the parallel-flow lines fan open at the inlet and close at the outlet, forming a wedge. The ratio is about 1.48, so counter flow gives roughly 48 percent more heat-transfer driving force — for the same Q you need about 1.48 times more area in parallel flow.

🙋

If counter flow is that much better, why do parallel-flow exchangers still exist in the field?

🎓

Three legitimate reasons. (1) Lower thermal shock: in counter flow the wall sees a large ΔT at one outlet, which can stress brittle materials; parallel flow has a small outlet ΔT and is gentler. (2) Reduced freezing or solidification risk: cooling molten plastics or wax counter-flow can over-cool and clog the outlet, parallel flow eases the slope. (3) Manufacturing convenience for some compact geometries. Still, the 30 to 40 percent loss in driving force means modern practice is counter flow or cross-flow by default. Drag T_h_out close to T_c_out in this tool and watch the parallel LMTD collapse toward invalid; meanwhile counter LMTD remains comfortable.

🙋

What does the bottom-right LMTD/ΔT_1 chart show? Why are there two dots?

🎓

Excellent. The LMTD formula is fully captured by a single dimensionless ratio r = ΔT_2/ΔT_1: LMTD/ΔT_1 = (1−r)/ln(1/r). That curve peaks at 1 when r = 1 (equal end differences) and drops fast as r approaches zero. Counter flow keeps both ΔTs close, so r is near 1 — the yellow dot sits high on the curve. Parallel flow has one big ΔT and one tiny ΔT, so r is small and the orange dot sits low. That single picture explains the 1.48 ratio at a glance. Move the sliders and watch both dots slide along the same universal curve.

Frequently Asked Questions

The Log Mean Temperature Difference is a representative temperature difference along the heat-transfer surface where the local ΔT between the hot and cold streams varies. It is given by LMTD = (ΔT_1 - ΔT_2) / ln(ΔT_1/ΔT_2), where ΔT_1 and ΔT_2 are the temperature differences at the two ends of the exchanger. Substituting it in Q = U A LMTD provides the total heat transfer rate integrated over the area. With the defaults T_h_in=150C, T_h_out=90C, T_c_in=25C, T_c_out=80C this tool reports counter-flow LMTD about 67.5 K and parallel-flow LMTD about 45.5 K.

In counter flow the hot and cold streams move in opposite directions, so a substantial temperature difference is maintained at both ends of the exchanger. In parallel flow the streams enter on the same side: ΔT is huge at the inlet but collapses to a small value at the outlet because both streams approach each other in temperature. The average driving force is larger in counter flow, so for the same U and A more heat is transferred, or equivalently less area is needed for the same Q. With the defaults this tool shows a counter / parallel LMTD ratio of about 1.48, a roughly 48 percent advantage for counter flow.

Heat always flows from hot to cold, so in parallel flow the hot outlet T_h_out can never become colder than the cold outlet T_c_out: this temperature crossover violates the second law of thermodynamics. If you push T_h_out below T_c_out in this tool the parallel-flow LMTD is flagged as invalid because ΔT_2 becomes non-positive and the logarithm is undefined. Counter-flow exchangers tolerate the milder regime where T_c_out can exceed T_h_out (true temperature crossover), which is part of why counter flow is thermally more efficient.

When ΔT_1 = ΔT_2 the expression (ΔT_1 - ΔT_2) / ln(ΔT_1/ΔT_2) is formally 0/0, but its limit is exactly LMTD = ΔT_1 = ΔT_2 by l'Hopital's rule. This tool returns ΔT_1 directly in that special case to avoid numerical singularities. Physically it corresponds to a constant temperature difference along the exchanger and occurs naturally in a balanced counter-flow exchanger where the heat-capacity rates of the two streams are equal (C* = 1). It is also the limit in which arithmetic and logarithmic means agree.

Real-World Applications

Shell-and-tube exchanger area sizing: In chemical plants and power stations the shell-and-tube exchanger is sized from the process-dictated Q and the four end temperatures via A = Q / (U LMTD F). F is a flow-arrangement correction factor (1.0 for pure counter flow, about 0.8 to 0.95 for a 1-2 shell-and-tube layout) tabulated in TEMA standards. Reading the counter LMTD from this tool with a typical liquid-liquid U of 1000 W/m²K and Q = 1 MW gives A = Q / (U LMTD) = 14.8 m² needed — a useful back-of-the-envelope sanity check before vendor selection.

HVAC chilled-water and hot-water coils: A building cooling coil moves the supply air from about 35°C to 14°C while chilled water enters at 7°C and leaves at 12°C. A parallel-flow coil would collapse to ΔT_2 = 14 − 12 = 2 K at the outlet; a counter-flow coil instead maintains ΔT_1 = 35 − 12 = 23 K and ΔT_2 = 14 − 7 = 7 K, giving LMTD ≈ 13.8 K. Entering those numbers in this tool confirms a roughly 2x driving force compared to the parallel arrangement, which is why HVAC coils are now built counter-flow by default.

Nuclear-plant steam generators: A PWR steam generator transfers heat from the primary coolant (300°C → 280°C) to boiling secondary water at 230°C. Entering T_h_in=300, T_h_out=280, T_c_in=230, T_c_out=230 (constant temperature during boiling) yields counter-flow LMTD ≈ 59 K and an area requirement of about 5,500 m² for 1 GW thermal at U = 3000 W/m²K. Real U-tube steam generators are a hybrid of cross flow and counter flow and use an F correction of about 0.9.

Economizers and waste-heat recovery: An industrial-furnace economizer recovers heat from exhaust gas (200°C → 100°C) into combustion air (25°C → 150°C). Parallel flow would require a temperature crossover (150 above 100) that the second law forbids, so the configuration is impossible. Counter flow is fine and yields LMTD ≈ 62 K. Entering those numbers in this tool flags parallel-flow as invalid while counter-flow remains valid, which is why nearly every waste-heat recovery unit is a counter-flow design.

Common Misconceptions and Caveats

The most common misconception is that "LMTD is just an arithmetic average of the two end ΔTs". In fact the arithmetic mean (ΔT_1 + ΔT_2)/2 is always larger than the LMTD, and the gap widens with the ΔT_1/ΔT_2 ratio. For the parallel-flow defaults of this tool (ΔT_1 = 125, ΔT_2 = 10) the arithmetic mean equals 67.5 K, while the LMTD is only 45.5 K — using the arithmetic mean would undersize the area by about 33 percent. In practice the arithmetic approximation is acceptable only for ΔT_1/ΔT_2 below about 2 (error under 4 percent); beyond that you must use the logarithmic form. Drag the T_h_out slider in this tool to see the gap widen.

Another frequent error is to assume "LMTD alone handles every heat-exchanger geometry". The pure LMTD result is exact only for ideal counter flow or parallel flow. For shell-and-tube 1-2 layouts, multi-pass arrangements and cross-flow exchangers the actual driving force is reduced by an F correction factor (less than 1): Q = U A F LMTD. F is read from TEMA or Bowman charts; designs typically require F above 0.75 (below that the configuration is unstable and a different geometry is preferred). This tool computes the F = 1 ideal — apply the correction separately for real hardware.

The third pitfall is to think that "LMTD and NTU-effectiveness are two unrelated methods". They are the same thermodynamics seen from two complementary angles. LMTD is convenient when both outlet temperatures are known and the unknown is the area; NTU-ε is convenient when the area and U are known and the unknowns are the outlets. The relation ε = f(NTU, C*) comes from the same energy balance as Q = U A LMTD. This tool is dedicated to LMTD — for the NTU-ε side use the companion heat-exchanger NTU simulator; together they cover both sizing and rating.

LMTD Simulator — Log Mean Temperature Difference for Counter and Parallel Flow

What is the LMTD Simulator?

Frequently Asked Questions

Real-World Applications

Common Misconceptions and Caveats

Related Tools