When you're sizing a cooling circuit or a hot runner manifold heater, you're really solving three heat transfer problems at once. Conduction governs heat moving through the mold steel itself: Q = k·A·ΔT/L, where k for P20 tool steel sits around 29 W/m·K, and for H13 it's closer to 24 W/m·K. That difference matters when you're calculating how fast heat escapes from the cavity surface to the cooling channel wall. In practice, most of us work with a simplified steady-state approach, but on thick cores or deep ribs, transient conduction often dominates the first few seconds of the cycle, so don't ignore thermal diffusivity when cycle time is tight.

Convection is where the cooling channel design actually earns its keep. The standard formula Q = h·A·ΔT lives or dies on the heat transfer coefficient h, which depends on coolant velocity, channel diameter, and whether you've got turbulent or laminar flow. Reynolds number above roughly 10,000 gives you turbulent flow and a dramatically higher h — often 3,000 to 6,000 W/m²·K for water in a 10 mm channel at 2 m/s. Drop to laminar flow and you can lose half that performance, which is why baffles, bubblers, and conformal cooling layouts exist. Radiation, by contrast, is usually a minor term in injection molds, but it becomes relevant in hot runner nozzles and high-temperature processes like PEEK or PPS, where Q = ε·σ·A·(T₁⁴ − T₂⁴) with emissivity values typically between 0.7 and 0.9 for oxidized steel surfaces.

The takeaway for the shop floor: calculate each mode separately, then check which one limits your cycle. If conduction through the steel is your bottleneck, no amount of extra coolant flow will fix it — you need to move the channel closer or switch to a higher-conductivity insert like beryllium copper. If convection is limiting, increase turbulence before you add more circuits. Getting this balance right is what separates a mold that runs at 18 seconds from one that runs at 25. For more mold sourcing and technical resources, visit MoldWorld at www.moldw.com.