Regenerative cooling — heat flux, coolant flow, pressure drop

1. Heat flux: how hard the wall is being cooked

WHAT drives it? Heat must travel: hot gas → wall (gas side) → conduction through metal → coolant. Each stage is a resistance in series, like resistors in an electrical circuit.

Deriving $q$ from first principles

Start from Newton's law of cooling at each interface and Fourier conduction in between. Define per-area (specific) resistances:

Gas-side convection: $R_g = 1/h_g$
Wall conduction: $R_w = t_w / k_w$ (thickness $t_w$ , conductivity $k_w$ )
Coolant-side convection: $R_c = 1/h_c$

HOW to combine: In steady state the same $q$ passes through each layer, so: $q = h_g\,(T_{aw}-T_{wg}) = \frac{k_w}{t_w}(T_{wg}-T_{wc}) = h_c\,(T_{wc}-T_{co})$

Why this step? Steady state means no heat piles up anywhere, so flux is conserved layer-to-layer. Add the temperature drops (they telescope):

$\boxed{q = \dfrac{T_{aw} - T_{co}}{\dfrac{1}{h_g} + \dfrac{t_w}{k_w} + \dfrac{1}{h_c}}}$

WHY $T_{aw}$ and not flame temperature? Near a wall the fast gas is slowed; friction reheats it partway, so the effective driving temperature is the recovery temperature $T_{aw}=T_c\left(\dfrac{1+r\frac{\gamma-1}{2}M^2}{1+\frac{\gamma-1}{2}M^2}\right)$ with recovery factor $r\approx \text{Pr}^{1/3}$ . Slightly below stagnation, but close to it.

2. Coolant flow: carrying the heat away

The channel absorbs total power $Q = q\,A$ over wetted area $A$ . The coolant heats up from inlet $T_{in}$ to outlet $T_{out}$ .

Why this step? Every joule entering the wall must be stored in the coolant's rising temperature (no phase change assumed). Energy in = energy carried out.

Constraint: $T_{out}$ must stay below the coolant's boiling/coking limit, otherwise film boiling insulates the wall → burnout. This caps $\Delta T_c$ , forcing $\dot m_c$ up.

The coolant-side heat transfer coefficient itself depends on flow via the Dittus–Boelter correlation: $\text{Nu}=0.023\,\text{Re}^{0.8}\text{Pr}^{0.4}, \quad h_c=\frac{\text{Nu}\,k_c}{d_h}$ WHY it matters: faster flow → higher Re → higher $h_c$ → lower wall temperature. More flow cools twice: bigger $\dot m_c$ AND bigger $h_c$ .

3. Pressure drop: the cost of thin channels

Pushing coolant through narrow, long channels costs pressure. That pressure must be supplied by the pumps — higher $\Delta p$ ⇒ heavier turbopump ⇒ performance penalty.

HOW derived (first principles): A pressure force $\Delta p\,A_{cs}$ pushes the fluid; wall shear $\tau_w$ resists over the wetted surface $P L$ . In steady flow they balance: $\Delta p\,A_{cs} = \tau_w\,P\,L \;\Rightarrow\; \Delta p = \tau_w\,\frac{PL}{A_{cs}} = \tau_w\,\frac{4L}{d_h}$ using $d_h = 4A_{cs}/P$ . Then define $f$ by $\tau_w = \tfrac{f}{4}\cdot\tfrac{\rho v^2}{2}$ (empirical dimensionless friction), giving the boxed form.

Why $v^2$ ? Because turbulent wall shear scales with dynamic pressure $\tfrac12\rho v^2$ — the momentum flux the wall must destroy.

Figure — Regenerative cooling — heat flux, coolant flow, pressure drop

Worked examples

Worked example Example 1 — Wall heat flux

Given $h_g=2.0\times10^4$ , $t_w=1$ mm, $k_w=350$ (copper), $h_c=5.0\times10^4$ W/m²K, $T_{aw}=3300$ K, $T_{co}=300$ K. Find $q$ and gas-side wall temp $T_{wg}$ .

Resistances: $R_g=1/2\times10^4=5.0\times10^{-5}$ ; $R_w=0.001/350=2.86\times10^{-6}$ ; $R_c=1/5\times10^4=2.0\times10^{-5}$ . Why? Sum the series specific resistances. Total $R=7.29\times10^{-5}$ m²K/W. $q=\frac{3300-300}{7.29\times10^{-5}} = 4.1\times10^{7}\ \text{W/m}^2 = 41\ \text{MW/m}^2$ Why huge? Rocket walls really do see tens of MW/m² — that's why cooling is life-or-death. $T_{wg}=T_{aw}-q R_g = 3300 - 4.1\times10^7(5\times10^{-5}) = 3300-2058 \approx 1242$ K. Below copper's melt — safe.

Common mistake Steel-manned misconceptions

(a) "Just make the wall thicker for safety." Feels right: thicker metal = stronger, more thermal mass. Reality: thicker wall raises $R_w=t_w/k_w$ , so $T_{wg}$ rises — hotter, not safer. Fix: rocket walls are made thin (≈1 mm) with high- $k$ copper alloys to keep $R_w$ tiny.

(b) "More coolant flow always helps, no downside." Feels right: more flow carries more heat and raises $h_c$ . Reality: $\Delta p\propto v^2$ , so doubling velocity quadruples pump work. Fix: optimize — enough flow for cooling, not so much you break the pump budget.

(c) "Use the flame temperature $T_c$ as the driving temperature." Feels right: that's the hottest number available. Reality: the gas is moving fast; only the recovery temperature $T_{aw}<T_c$ drives heat into the wall. Fix: use $T_{aw}=r$ -corrected value; overestimating $q$ wastes coolant.

Recall Feynman: explain to a 12-year-old

Imagine a metal cup so hot it would melt — like next to a blowtorch. Before you drink your cold juice, you first run the juice through little tubes wrapped around the cup. The cold juice grabs the heat, keeping the cup cool. Then you drink the (now warm) juice — nothing wasted! But if the tubes are too skinny, you have to suck really hard to get the juice through. That "sucking hard" is the pressure drop, and it's the annoying cost.

Recall Active recall — cover the answers

What resistances add to give $1/U$ ? ::: gas convection $1/h_g$ , wall conduction $t_w/k_w$ , coolant convection $1/h_c$ .
Why thin walls? ::: to minimize $R_w=t_w/k_w$ so wall stays cool.
Why does more velocity cool better but cost more? ::: $h_c\uparrow$ (Re $^{0.8}$ ) but $\Delta p\propto v^2$ .

Flashcards

What is heat flux and its units?

Thermal power per unit wall area, W/m². It's a power density, not total power.

Write the series-resistance formula for overall heat flux.

q=(T_{aw}-T_{co})/(1/h_g + t_w/k_w + 1/h_c)

Why use recovery temperature

T_{aw}

instead of flame temperature?

The fast gas is only partly reheated near the wall; the effective driving temp is

T_{aw}=r

-corrected, slightly below stagnation.

Why does thickening the wall make it hotter, not safer?

It increases

R_w=t_w/k_w

, raising the gas-side wall temperature

T_{wg}=T_{aw}-qR_g

... actually raises

T_{wc}

; overall higher wall temps.

How do you compute required coolant mass flow?

\dot m_c=qA/(c_p\Delta T_c)

from energy balance

Q=\dot m_c c_p\Delta T_c

State Darcy–Weisbach and derive its origin.

\Delta p=f\frac{L}{d_h}\frac{\rho v^2}{2}

; from pressure force = wall shear force,

\Delta p A_{cs}=\tau_w P L

, with

d_h=4A_{cs}/P

What limits the allowed coolant temperature rise?

Boiling/coking of coolant; film boiling would insulate and cause wall burnout.

Why is regenerative cooling called "regenerative"?

Heat absorbed by coolant is returned to the combustion when that propellant is burned — energy is recovered.

How does

h_c

depend on flow (Dittus–Boelter)?

Nu=0.023Re^{0.8}Pr^{0.4}

, so

h_c\propto Re^{0.8}\propto v^{0.8}

The central design trade-off in one line?

Smaller channels cool better (

h_c\uparrow

) but cost more pump work (

\Delta p\propto v^2/d_h

Connections

Newton's Law of Cooling — gas- and coolant-side convection terms.
Fourier's Law of Conduction — the $t_w/k_w$ wall term.
Dittus-Boelter Correlation — sets $h_c$ from Reynolds & Prandtl.
Darcy-Weisbach Equation — pressure drop in the coolant channels.
Adiabatic Wall Temperature and Recovery Factor — the real driving temperature.
Turbopump Sizing — where $\Delta p$ becomes a mass penalty.
Film Cooling & Ablative Cooling — alternative wall-protection schemes.

Concept Map

Hinglish (regional understanding)

Intuition Hinglish mein samjho

Dekho, rocket engine ke andar gas ka temperature 3000+ K hota hai — koi bhi metal seedha melt ho jaaye. To trick kya hai? Propellant ko burn karne se pehle, usko wall ke andar bani choti-choti channels me se pump karte hain. Thanda propellant heat ko soak kar leta hai, wall thanda rehta hai, aur woh heat waste nahi hoti — kyunki wahi propellant baad me jal jaata hai. Isiliye naam hai regenerative cooling.

Teen cheezein important hain. Pehla heat flux $q$ — kitni tezi se heat wall me ghus rahi hai (W/m²). Isko nikalte hain series resistance se: gas-side film $1/h_g$ , metal conduction $t_w/k_w$ , aur coolant film $1/h_c$ — teenon add hote hain jaise circuit me resistors. Formula: $q=(T_{aw}-T_{co})/(1/h_g+t_w/k_w+1/h_c)$ . Yaad rakho, driving temperature flame temp nahi, balki recovery temperature $T_{aw}$ hoti hai, thodi kam.

Doosra coolant flow $\dot m_c$ — energy balance se: total power $Q=qA$ ko coolant ka temperature rise nikaalta hai, $\dot m_c=qA/(c_p\Delta T_c)$ . Coolant ka temperature bahut nahi badhna chahiye warna boiling ho jaayegi aur wall jal jaayega. Teesra pressure drop $\Delta p=f\frac{L}{d_h}\frac{\rho v^2}{2}$ — patli channel me tez flow behtar cooling deta hai par pump ko zyada mehnat karni padti hai.

Asli maza yahi trade-off hai: channel patli karo to $h_c$ badhta hai (better cooling), par $\Delta p$ velocity ke square se badhta hai (heavy turbopump). Engineer ka kaam — itna hi cool karo ki wall bache, par itna tight na karo ki pump hi fail ho jaaye. Yeh 20% samajh lo, poora regen cooling clear ho jaayega.

Regenerative cooling — heat flux, coolant flow, pressure drop

1. Heat flux: how hard the wall is being cooked

Deriving $q$ from first principles

2. Coolant flow: carrying the heat away

3. Pressure drop: the cost of thin channels

Worked examples

Flashcards

Connections

Concept Map

Hinglish (regional understanding)

Go deeper — visual, from zero

Test yourself — Rocket Propulsion

Connections