Designing a heat exchanger starts with defining your thermal duty: how much heat needs to transfer between two fluids, at what temperatures, and under what constraints. From there, the process moves through selecting a configuration, sizing the heat transfer surface, choosing materials, and verifying that pressure drops stay within acceptable limits. Each step feeds into the next, and getting the sequence right saves you from costly iterations later.
Define the Thermal Duty First
Before anything else, pin down the basic parameters of your problem. You need the mass flow rates of both fluids, their inlet temperatures, the desired outlet temperatures (for at least one fluid), and the specific heat capacities of each fluid. The fundamental energy balance is straightforward: the heat released by the hot fluid equals the heat absorbed by the cold fluid, minus any losses. This gives you Q, the total heat transfer rate in watts, which drives every downstream calculation.
If you only know one outlet temperature, you can calculate the other from the energy balance. If you know neither outlet temperature, that changes which sizing method you’ll use, as discussed below.
Choose the Right Configuration
The type of heat exchanger you select depends on the fluids involved, the operating pressures, the required surface area, and space constraints. The most common configurations fall into a few categories:
- Shell-and-tube: The workhorse of industrial applications. One fluid flows through a bundle of tubes while the other flows around them inside a cylindrical shell. Handles high pressures, is easy to clean, and scales well. Multiple tube passes increase heat transfer at the cost of higher pressure drop.
- Plate heat exchangers: Thin corrugated plates stacked together with alternating fluid channels. Compact, efficient, and excellent for moderate pressures. Common in HVAC, food processing, and chemical applications where you need a lot of surface area in a small footprint.
- Double-pipe: The simplest design, with one pipe inside another. Best for small duties or as a starting point for learning the fundamentals. Limited surface area makes it impractical for large-scale applications.
- Cross-flow: One fluid flows perpendicular to the other, typical in air-cooled systems like car radiators. Useful when one fluid is a gas with much lower heat transfer properties than the liquid on the other side.
Within any configuration, you also choose between parallel flow (both fluids entering at the same end) and counterflow (fluids entering at opposite ends). Counterflow is almost always more thermally efficient because it maintains a more uniform temperature difference along the length of the exchanger, allowing you to achieve closer approach temperatures with less surface area.
Size the Exchanger With LMTD or Effectiveness-NTU
Two methods dominate heat exchanger sizing, and which one you use depends on what information you have at the start.
The LMTD Method
The Log Mean Temperature Difference method works best when you know the inlet and outlet temperatures of both fluids. You calculate the LMTD, which represents the effective average temperature driving force across the exchanger, then plug it into the core equation:
Q = U × A × LMTD
Here, U is the overall heat transfer coefficient (combining the resistances of both fluid films and the tube wall), A is the heat transfer surface area, and LMTD accounts for the fact that the temperature difference between the two fluids changes along the exchanger’s length. When you know Q, U, and LMTD, solving for A gives you the required surface area directly.
For anything other than simple parallel or counterflow arrangements, you need a correction factor (commonly called the F-factor) that adjusts the LMTD downward. A multipass shell-and-tube exchanger, for instance, doesn’t behave like a pure counterflow unit. The corrected equation becomes Q = U × A × LMTD × F. The F-factor depends on two dimensionless ratios: one capturing how much the tube-side fluid temperature changes relative to the total temperature span, and another representing the ratio of heat capacities between the two streams. Published charts for common configurations let you look up F once you calculate these ratios. In a worked example from EPFL, a two-shell-pass, four-tube-pass oil cooler with oil entering at 181°C and leaving at 38°C, and water entering at 32°C and leaving at 49°C, produces an LMTD of about 41 K and an F-factor of 0.92, yielding a required area of 121 square meters.
The limitation: when you don’t know the outlet temperatures of both fluids, you can’t calculate LMTD directly. You’re forced into iteration, guessing outlet temperatures, checking the energy balance, and repeating until everything converges. This is where the second method shines.
The Effectiveness-NTU Method
When only the inlet temperatures are known (a common situation when you’re rating an existing exchanger or exploring different operating conditions), the effectiveness-NTU method avoids iteration entirely. It defines effectiveness as the ratio of actual heat transfer to the maximum possible heat transfer. The Number of Transfer Units (NTU) is a dimensionless measure of the exchanger’s size relative to the flow conditions: NTU = UA / C_min, where C_min is the smaller of the two fluids’ heat capacity rates.
Each exchanger configuration has its own effectiveness-NTU relationship, available as equations or charts. You calculate NTU from known geometry and flow conditions, determine the effectiveness, then multiply by the maximum possible Q to get the actual heat duty. This method is especially useful for comparing how an existing exchanger performs under different flow rates or inlet temperatures.
Calculate the Heat Transfer Coefficient
The overall heat transfer coefficient U is the single most important (and most uncertain) number in your design. It combines three resistances in series: convection on the hot side, conduction through the tube wall, and convection on the cold side. Fouling on either surface adds additional resistance that you must account for with fouling factors, which are based on the type of fluid and operating conditions.
For the convective portions, you need to estimate the heat transfer coefficient on each side. In turbulent flow through tubes (Reynolds number above 10,000), the most widely used correlation relates the Nusselt number to the Reynolds and Prandtl numbers:
Nu = 0.023 × Re^0.8 × Pr^n
The exponent n equals 0.4 when the fluid is being heated and 0.3 when it’s being cooled. This correlation is valid for Prandtl numbers between 0.7 and 160 and for tubes where the length-to-diameter ratio is at least 10. Once you have the Nusselt number, you extract the convective heat transfer coefficient by multiplying by the fluid’s thermal conductivity and dividing by the tube diameter.
When fluid properties vary significantly between the bulk temperature and the wall temperature (common with viscous oils), alternative correlations include a viscosity correction term that accounts for this difference. The key point is that your U value is only as good as your estimates of these individual coefficients, so using the right correlation for your flow regime matters.
For laminar flow (Reynolds number below 2,300), heat transfer is much lower and different correlations apply. If possible, designing for turbulent flow on both sides improves performance significantly, though it increases pressure drop.
Select Materials for Tubes and Shell
Material choice balances thermal performance, corrosion resistance, mechanical strength, and cost. Thermal conductivity determines how easily heat passes through the tube wall itself. Copper leads the pack at 385 W/m·K, making it the default choice for clean, non-corrosive applications like HVAC condensers and domestic hot water systems. Steel comes in at roughly 50 W/m·K, about one-eighth of copper’s conductivity, but offers better mechanical strength and lower cost for large industrial units. Stainless steel and titanium sacrifice some thermal conductivity for dramatically better corrosion resistance in aggressive chemical environments or seawater service.
In practice, the tube wall resistance is often a small fraction of the total thermal resistance. The fluid-side convective resistances and fouling layers dominate, which means upgrading from steel to copper tubes sometimes yields less improvement than you’d expect. Focus material selection on ensuring the exchanger survives its operating environment over its intended lifespan. A cheaper material that corrodes and fouls rapidly will underperform a more expensive one that stays clean.
Check Pressure Drop
Every heat exchanger design involves a tradeoff between heat transfer and pressure drop. Higher fluid velocities improve convective heat transfer but increase the energy needed to push fluid through the system. On the tube side, pressure drop depends on tube length, diameter, number of passes, and flow velocity. On the shell side, the baffle spacing and arrangement control how the fluid moves across the tube bundle.
Your design is constrained by the available pumping power. A beautifully efficient thermal design is useless if the pressure drop exceeds what your pumps can deliver. As a practical matter, calculate pressure drops on both sides after your initial thermal sizing, then adjust geometry if needed. Reducing the number of tube passes, increasing tube diameter, or widening baffle spacing all reduce pressure drop at the expense of some thermal performance.
Iterate and Verify
Heat exchanger design is inherently iterative. Your first pass through the calculations gives you a starting geometry. You then check whether the assumed U value matches what the geometry and flow conditions actually produce. If it doesn’t, you adjust the design (more tubes, longer tubes, different baffle spacing) and recalculate. Most designers go through three to five iterations before the assumed and calculated U values converge within a few percent.
Once converged, add margin. Fouling will degrade performance over time, and real-world operating conditions rarely match design conditions exactly. A common approach is to oversize the surface area by 10 to 25 percent, depending on how fouling-prone the fluids are. Too much margin, though, creates problems of its own: an oversized exchanger at startup may subcool or overheat one of the streams, requiring control measures like bypass valves.
Software tools like HTRI, Aspen Exchanger Design, and even well-built spreadsheets handle the iterative math efficiently, but understanding the underlying method lets you interpret results critically and catch errors that software won’t flag on its own.