Heat exchangers: LMTD method
Log Mean Temperature Difference (LMTD) method for sizing and rating heat exchangers, including counterflow/parallel-flow cases, correction factor F for multipass/crossflow, assumptions, derivation, and exam-focused applications.
Drafted with Aria, reviewed by the AiCanCode.org team · Updated . How we write and check our notes. Spotted an error? Use Give Feedback at the bottom of the page.
Why it matters
Heat exchangers are everywhere: power plants, chemical processes, HVAC, automotive, and electronics cooling. The LMTD method is the fastest way to size a heat exchanger when the inlet and outlet temperatures are specified or can be assumed, making it central to design calculations and quick checks in industry and exams.
Key ideas
- A heat exchanger transfers heat between two fluids separated by a solid wall. We treat this topic; for fluid properties and convection coefficients, see Convection Heat Transfer; for wall conduction, see Conduction Heat Transfer; for radiation, see Radiation Heat Transfer.
- Overall heat transfer rate (Q) links thermal driving force and resistance: heat flows from hot to cold due to a temperature difference, resisted by convection on both sides, wall conduction, and fouling.
- Log Mean Temperature Difference (LMTD) is the effective average temperature difference between the fluids along the exchanger when U is roughly constant. It replaces a variable ΔT(x) with a single equivalent ΔT_lm that gives the same Q when multiplied by U·A.
- Exact statement (single-pass, steady, no losses): For a given flow arrangement,
Q = U·A·ΔT_lm, whereΔT_lm = (ΔT1 − ΔT2) / ln(ΔT1/ΔT2), with ΔT1 and ΔT2 the terminal hot–cold differences at the two ends of the exchanger. - Flow arrangements:
- Parallel flow: both fluids enter at the same end; ΔT1 = (T_h,in − T_c,in); ΔT2 = (T_h,out − T_c,out). The temperature difference decreases along the length; cold outlet cannot exceed the hot outlet temperature.
- Counterflow (usually gives higher ΔT_lm): fluids enter at opposite ends; ΔT1 = (T_h,in − T_c,out); ΔT2 = (T_h,out − T_c,in). A temperature “cross” is possible: the cold outlet may exceed the hot outlet temperature; never exceeds the hot inlet temperature for sensible heating/cooling.
- Assumptions in classic LMTD derivation:
- Steady state; no heat loss to surroundings (adiabatic shell); negligible axial conduction in fluids and wall; constant overall heat transfer coefficient U; constant cp; no phase change (single-phase) unless treated as a special isothermal side.
- Uniform flow distribution; no bypassing or leakage.
- Derivation outline (single-pass):
- Consider a differential area dA. Local heat flux: dQ = U·(T_h − T_c)·dA.
- Energy balances: dQ = −C_h·dT_h = +C_c·dT_c, where C_h = ṁ_h·c_p,h and C_c = ṁ_c·c_p,c.
- Eliminate dQ and relate d(T_h − T_c) to dA. Integrate from inlet to outlet along the flow path respecting the arrangement (parallel or counterflow).
- Result: Q = U·A·(ΔT1 − ΔT2)/ln(ΔT1/ΔT2), with ΔT1 and ΔT2 defined per arrangement above. Limiting case when ΔT1 = ΔT2 gives ΔT_lm = ΔT1.
- Correction factor (F) for multipass and crossflow exchangers:
- Real shell-and-tube with multiple tube passes or one shell pass (1–2, 2–4, etc.) and crossflow exchangers do not see a pure counterflow/parallel pattern. We use:
Q = U·A·F·ΔT_lm,cf, where ΔT_lm,cf is the counterflow LMTD computed from terminal temperatures and F ≤ 1 corrects for deviation from true counterflow. - F depends on the geometry and on two dimensionless temperature ratios: R = (T_h,in − T_h,out)/(T_c,out − T_c,in) and **P = (T_c,out − T_c,in)/(T_h,in − T_c,in)`. F is read from standard F-charts for the specific exchanger type (e.g., 1–2 shell-and-tube) using R and P. As a design guideline, F ≥ 0.75 is commonly recommended; very low F indicates poor temperature driving-force distribution (reconsider passes or area).
- Real shell-and-tube with multiple tube passes or one shell pass (1–2, 2–4, etc.) and crossflow exchangers do not see a pure counterflow/parallel pattern. We use:
- Overall heat transfer coefficient (U):
- Defined by
Q = U·A·ΔT_lm(orU·A·F·ΔT_lm,cf). U lumps convection, conduction, and fouling: 1/(U·A_base) = Σ thermal resistances referred to A_base (commonly tube outside area A_o or inside area A_i — be consistent). - For a typical tube wall, referred to outside area A_o: 1/(U_o·A_o) = 1/(h_o·A_o) + (ln(r_o/r_i))/(2πkL) + 1/(h_i·A_i) + R_f,o + R_f,i (with appropriate area conversion for fouling resistances). See Heat Exchangers and Conduction Heat Transfer topics for full resistance networks.
- Defined by
- Standard and limiting cases:
- If one fluid undergoes phase change (condensation/boiling) at constant temperature, the ΔT at that side’s end is simply the difference between the fixed phase-change temperature and the other fluid’s end temperature; the same LMTD formula applies with those ΔTs.
- When ΔT1 ≈ ΔT2, ΔT_lm ≈ their arithmetic mean; exactly equal implies ΔT_lm equals that common value (use the limit of the formula).
- Counterflow always yields ΔT_lm ≥ parallel flow for the same terminal temperatures; thus, for given U·A, counterflow achieves more heat transfer (or needs less area).
- Why real devices differ from ideal:
- U varies along length (property and Re changes), maldistribution reduces effective area, bypass/leakage in shells reduces true crossflow, fouling increases resistance over time, and there may be heat loss to surroundings. These effects lower effective U·A and shift terminal temperatures from ideal predictions.
- When to use LMTD vs. NTU (see Heat Exchangers: NTU method):
- LMTD: Best when all four terminal temperatures are known (rating) or two inlets and two outlets are specified by process needs (sizing). If an outlet is unknown, iteration is required.
- NTU–effectiveness: Convenient when only inlets are known and U·A is specified; avoids iteration to get outlets. In design practice both methods cross-check each other.
- Sign conventions and feasibility:
- Take Q positive from hot to cold: Q = C_h(Th,in − Th,out) = C_c(Tc,out − Tc,in) > 0 for cooling of hot fluid and heating of cold fluid.
- Parallel flow constraint: Tc,out < Th,out for sensible heating/cooling. Counterflow allows Tc,out > Th,out, but Tc,out < Th,in (no heat pumps here).
Formulas
Q = U·A·ΔT_lm — Heat transfer rate for single-pass ideal arrangements; Q in W, U in W·m⁻²·K⁻¹, A in m², ΔT_lm in K. Applies to parallel or counterflow with ΔT_lm computed for that arrangement; assumes steady state, no losses, constant U.
ΔT_lm = (ΔT1 − ΔT2) / ln(ΔT1/ΔT2) — Log mean temperature difference; ΔT1, ΔT2 in K are the terminal hot–cold differences at the two ends (defined by flow pattern). Valid for monotonic temperature profiles without sign change.
ΔT1 = Th,in − Tc,out; ΔT2 = Th,out − Tc,in — Terminal temperature differences for counterflow; temperatures in °C or K (differences in K). Applies when both fluids are single-phase (or one phase-change with constant temperature) and counterflow.
ΔT1 = Th,in − Tc,in; ΔT2 = Th,out − Tc,out — Terminal differences for parallel flow; same units/assumptions as above.
Q = U·A·F·ΔT_lm,cf — With correction factor F for multipass/crossflow; ΔT_lm,cf is the counterflow LMTD from terminal temperatures; F ≤ 1 accounts for deviation from true counterflow. Geometry-specific F-charts give F as a function of R and P.
R = (Th,in − Th,out) / (Tc,out − Tc,in) — Temperature ratio for F-charts (dimensionless). Use with specified exchanger configuration (e.g., 1–2 shell-and-tube).
P = (Tc,out − Tc,in) / (Th,in − Tc,in) — Temperature effectiveness parameter for F-charts (dimensionless). Use with R to read F from charts.
Q = C_h·(Th,in − Th,out) = C_c·(Tc,out − Tc,in) — Energy balance; C_h = ṁ_h·c_p,h, C_c = ṁ_c·c_p,c in W·K⁻¹; valid for single-phase sensible heat transfer, no losses.
1 / (U_o·A_o) = 1 / (h_o·A_o) + ln(r_o/r_i) / (2πkL) + 1 / (h_i·A_i) + R_f,o + R_f,i — Overall resistance referred to outside area A_o for a cylindrical tube; h in W·m⁻²·K⁻¹, k in W·m⁻¹·K⁻¹, r in m, L in m, fouling resistances R_f in m²·K·W⁻¹. Applies to clean/fouled shell-and-tube; assumes uniform h and k along L.
Worked examples
Example 1 — Direct sizing with counterflow LMTD Given: Hot water cooled from Th,in = 90 °C to Th,out = 60 °C at ṁ_h = 1.5 kg·s⁻¹; Cold water heated from Tc,in = 25 °C at ṁ_c = 2.0 kg·s⁻¹; Assume c_p,water = 4180 J·kg⁻¹·K⁻¹; Overall U = 700 W·m⁻²·K⁻¹; Counterflow; Neglect heat loss.
Find: Required heat transfer area A.
Heat duty from hot side: Formula: Q = ṁ_h·c_p·(Th,in − Th,out) Numbers: Q = 1.5 kg·s⁻¹ × 4180 J·kg⁻¹·K⁻¹ × (90 − 60) K = 1.5 × 4180 × 30 W = 188,100 W
Check cold outlet temperature (consistency): Formula: Tc,out = Tc,in + Q/(ṁ_c·c_p) Numbers: Tc,out = 25 °C + 188,100 W / (2.0 × 4180 W·K⁻¹) = 25 + 22.5 = 47.0 °C (Energy balance satisfied within rounding.)
Terminal temperature differences for counterflow: Formulae: ΔT1 = Th,in − Tc,out; ΔT2 = Th,out − Tc,in Numbers: ΔT1 = 90 − 47.0 = 43.0 K; ΔT2 = 60 − 25 = 35.0 K
LMTD: Formula: ΔT_lm = (ΔT1 − ΔT2)/ln(ΔT1/ΔT2) Numbers: ΔT_lm = (43.0 − 35.0) K / ln(43.0/35.0) = 8.0 / ln(1.22857) = 8.0 / 0.205 = 39.0 K (approx.)
Area: Formula: A = Q / (U·ΔT_lm) Numbers: A = 188,100 W / (700 W·m⁻²·K⁻¹ × 39.0 K) = 188,100 / 27,300 = 6.89 m² (approx.)
Answer: A ≈ 6.9 m².
Assumptions: Constant U, negligible heat losses, water properties ~constant over this range.
Example 2 — GATE-level: Multipass shell-and-tube with correction factor and iteration Given: 1-shell-pass, 2-tube-pass (1–2) exchanger; Hot oil (c_p,h = 2200 J·kg⁻¹·K⁻¹) at Th,in = 150 °C, ṁ_h = 2.5 kg·s⁻¹; Cold water (c_p,c = 4180 J·kg⁻¹·K⁻¹) at Tc,in = 25 °C, ṁ_c = 3.0 kg·s⁻¹; Overall U = 600 W·m⁻²·K⁻¹; Area A = 25 m². Use correction factor F = 0.85 for this 1–2 exchanger (from standard F-charts for values of R and P near those obtained below). Steady state; negligible losses; single-phase.
Find: Outlet temperatures Th,out and Tc,out.
Pre-compute capacity rates:
- C_h = ṁ_h·c_p,h = 2.5 × 2200 = 5500 W·K⁻¹
- C_c = ṁ_c·c_p,c = 3.0 × 4180 = 12,540 W·K⁻¹
- U·A·F = 600 × 25 × 0.85 = 12,750 W·K⁻¹
Unknowns: Th,out and Tc,out. Equations: (1) Q = C_h(Th,in − Th,out) = C_c(Tc,out − Tc,in) (2) Q = (U·A·F)·ΔT_lm,cf with ΔT1 = Th,in − Tc,out; ΔT2 = Th,out − Tc,in We solve by iteration on Th,out.
Iteration 1 (guess Th,out = 60 °C):
- Q (energy) = 5500 × (150 − 60) = 495,000 W
- Tc,out = 25 + 495,000 / 12,540 = 64.51 °C
- ΔT1 = 150 − 64.51 = 85.49 K; ΔT2 = 60 − 25 = 35.00 K
- ΔT_lm = (85.49 − 35.00)/ln(85.49/35.00) = 50.49/0.892 = 56.6 K
- Q (LMTD) = 12,750 × 56.6 = 721,650 W (too high ⇒ need larger Q from energy side ⇒ lower Th,out)
Iteration 2 (Th,out = 50 °C):
- Q (energy) = 5500 × 100 = 550,000 W
- Tc,out = 25 + 550,000 / 12,540 = 68.86 °C
- ΔT1 = 150 − 68.86 = 81.14 K; ΔT2 = 50 − 25 = 25.00 K
- ΔT_lm = (81.14 − 25.00)/ln(81.14/25.00) = 56.14/1.177 = 47.68 K
- Q (LMTD) = 12,750 × 47.68 = 607,920 W (still high ⇒ Th,out slightly higher than needed)
Iteration 3 (Th,out = 47 °C):
- Q (energy) = 5500 × 103 = 566,500 W
- Tc,out = 25 + 566,500 / 12,540 = 70.17 °C
- ΔT1 = 150 − 70.17 = 79.83 K; ΔT2 = 47 − 25 = 22.00 K
- ΔT_lm = (79.83 − 22.00)/ln(79.83/22.00) = 57.83/1.288 = 44.90 K
- Q (LMTD) = 12,750 × 44.90 = 572,475 W (slightly higher than 566,500 W)
Refinement (Th,out = 46.8 °C):
- Q (energy) = 5500 × (150 − 46.8) = 5500 × 103.2 = 567,600 W
- Tc,out = 25 + 567,600 / 12,540 = 70.28 °C
- ΔT1 = 150 − 70.28 = 79.72 K; ΔT2 = 46.8 − 25 = 21.80 K
- ΔT_lm = (79.72 − 21.80)/ln(79.72/21.80) = 57.92/1.298 = 44.64 K
- Q (LMTD) = 12,750 × 44.64 = 569,160 W (matches within ~0.3%)
Accept:
- Th,out ≈ 46.8 °C
- Tc,out ≈ 70.3 °C
Check R and P (for information/consistency with the assumed F):
- R = (Th,in − Th,out)/(Tc,out − Tc,in) = (150 − 46.8)/(70.28 − 25) = 103.2/45.28 ≈ 2.28
- P = (Tc,out − Tc,in)/(Th,in − Tc,in) = 45.28/125 ≈ 0.362
- For a 1–2 exchanger, F-charts generally give F around 0.85 at these R and P. Our assumed F = 0.85 is consistent.
Answer: Th,out ≈ 46.8 °C; Tc,out ≈ 70.3 °C.
Notes: If no F-value were given, you would estimate R and P from a provisional iteration (as above), read F from standard charts, and iterate again including the updated F.
Common mistakes
- Using arithmetic mean ΔT instead of LMTD when ΔT varies significantly along length.
- Taking the logarithm of a ratio < 0 (e.g., when ΔT1 and ΔT2 signs differ or you mix up ends); LMTD needs both terminal differences positive and defined at the correct ends.
- Forgetting the correction factor F for multipass/crossflow, or assuming F = 1.
- Mixing base areas when computing U·A (e.g., U_o with A_i). Keep U and A consistent (both referred to inside or outside area).
- Ignoring energy balance: computed Q from LMTD must match ṁ·c_p·ΔT on both sides within acceptable tolerance.
- Using unrealistic constant U despite large property/flow changes; U may vary, particularly with large temperature spans or phase change on one side.
- Violating feasibility: predicting Tc,out > Th,out in parallel flow or Tc,out > Th,in in any sensible exchanger.
For GATE ME
- Compute ΔT_lm for parallel vs. counterflow given terminal temperatures; compare which arrangement needs less area.
- Size/rate a heat exchanger using Q = U·A·ΔT_lm (or U·A·F·ΔT_lm,cf) with given U, A, and temperatures.
- Apply correction factor F for a 1–2 shell-and-tube or crossflow; read F from provided chart using R and P or use a given F.
- Iterative outlet temperature calculation when one or both outlets are unknown; enforce energy balance and LMTD simultaneously.
- Conceptual: feasibility of temperature cross, effect of increasing number of shell/tube passes on F and area, why counterflow is thermally more effective than parallel.
- Impact of fouling on U and required area; adding fouling resistances to 1/U.
- What to practise:
- Derive LMTD quickly and remember terminal difference definitions for both flow patterns.
- Fast, accurate calculation of ΔT_lm and careful log handling; recognize limiting case ΔT1 ≈ ΔT2.
- Energy balance checks; quick iterations or structured trial for unknown outlets.
- Use and interpretation of F-charts (be fluent with R and P definitions and reading F when charts are provided).
- Estimating U·A effects and feasibility checks (temperature cross logic).
Quick check
- In counterflow, hot fluid cools from 120 °C to 80 °C; cold fluid heats from 30 °C to 70 °C. Write ΔT1 and ΔT2 for LMTD.
- A 1–2 shell-and-tube exchanger has R = 1.8 and P = 0.3. Is F likely closer to 0.6 or 0.9?
- If ΔT1 = ΔT2 = 25 K, what is ΔT_lm?
Answers: 1) ΔT1 = Th,in − Tc,out = 120 − 70 = 50 K; ΔT2 = Th,out − Tc,in = 80 − 30 = 50 K. 2) Closer to 0.9 (typical 1–2 charts give high F at moderate R and P). 3) ΔT_lm = 25 K.
Finished this topic? Mark it so your progress, study plan and readiness keep up.
Stuck on something here?