Modeling, Simulation & Computation · 270

Folding a Reactor's Cooling Water into One Bathtub — Building a Simple RCS Model

Heat in minus heat out, divided by the size of the container: a pencil-and-paper model of the reactor coolant system.

Series 270

A transparent glass bathtub rimmed with steel pipes, a metal tap pouring a glowing stream into swirling blue water, and a drain at the bottom with a whirlpool — A reactor coolant system folded into one bathtub: heat pours in from the tap, the steam generators drain it away. Conceptual illustration.

Key idea

Treat all the reactor's cooling water as one lump with a single temperature, and one energy balance tells you how fast it heats up: M·cp·dT/dt = Q_core − Q_SG. For an APR1400-class example, a 5% drop in steam-generator heat removal raises the average coolant temperature by about 0.14 °C per second, a little over 8 °C a minute, if nothing else responds. [1]

A puzzle you can solve on paper

At a large nuclear plant, the steam generators suddenly remove 5% less heat than usual, while the reactor keeps producing heat as before. How fast does the circulating coolant warm up? No supercomputer needed: the answer is a little over 8 degrees a minute, and the tool that gets you there is the simple RCS model.

Erase everything except the tap and the drain

The Reactor Coolant System (RCS) is a giant loop: water heated in the reactor flows to the steam generators, hands over its heat, and is pumped back. Picture it as a bathtub. The tap is the heat from the reactor core, the drain is the heat the steam generators remove, and the water level is the average coolant temperature.

Then make a bold simplification: treat all the water in the reactor, pipes and steam generators as one lump with a single temperature. This is a lumped model, the opposite end of the scale from computer codes that split the system into hundreds of cells.

Diagram of the primary coolant folded into one glass tank: 3,983 MW of core heat flowing in, steam-generator heat removal flowing out and dropping from 3,983 to 3,784 MW, a coolant lump of an assumed 250 tonnes at an average 307 degrees Celsius, and a card with three equations: the energy balance, delta T of about 33 degrees, and a time constant of about 8 seconds — Example values for an APR1400-class plant. The 250-tonne coolant mass is an educational assumption, not a design value.

One equation; the tub's floor area is the key

M·cp·dT/dt = Q_core − Q_SG

dT/dt is how fast the average temperature changes; the right side is heat in minus heat out. M is the coolant mass and cp its specific heat, so M·cp is the bathtub's floor area: the bigger the container holding the heat, the more slowly the temperature moves. [3]

Quantity	Example value
Thermal power	3,983 MW
Pressure	15.5 MPa
Specific heat at about 300 °C	about 5.7 kJ/(kg·°C) (room-temperature water: about 4.2)
Coolant mass M	250 t (assumed)
M·cp	about 1.43 GJ/°C
Imbalance after a 5% cut	about 199 MW
Rate of rise	0.14 °C/s, about 8.4 °C/min

A bonus: the core temperature rise is ΔT = Q/(ṁ·cp). With an example flow of about 21,000 kg/s, ΔT is about 33 °C, matching the APR1400's cold leg at about 291 °C and hot leg at about 324 °C, for an average of about 307 °C. [2]

Does the drain know the water level?

Whether the temperature climbs forever depends on the drain. A drain with a pump that always removes the same amount lets the tub fill until it overflows: fixed heat removal. An ordinary drain flows faster as the level rises, so the tub settles at a new level. A steam generator behaves like the second kind, removing more heat as the coolant gets hotter: Q_SG = UA·(Tavg − Ts), with the steam-side temperature Ts assumed to be about 285 °C.

Graph of average coolant temperature from 10 seconds before to 60 seconds after a 5 percent drop in steam-generator heat removal: both cases start with the same slope of 0.14 degrees per second, the fixed-removal case rises in a straight line by 8.4 degrees at 60 seconds, and the proportional case levels off 1.2 degrees higher with a time constant of about 8 seconds, beside a bathtub image and a list of what the model leaves out — Euler integration of the simple model at 0.01-second steps, with core power fixed at 3,983 MW, an assumed 250-tonne coolant mass and an assumed, fixed 285 °C steam-side temperature.

Both curves leave with the same slope. The fixed case rises about 8.4 °C in 60 seconds and keeps going; the proportional case levels off about 1.2 °C higher. The time constant τ = M·cp/UA, about 8 seconds here, is the time to cover 63% of the final rise, not the time to finish; nearly catching up takes three or four times longer.

What this bathtub leaves out

Reactivity feedback: a pressurized water reactor lowers its own power when the coolant heats up, so the real rise is gentler than with Q_core held fixed.
The pressurizer: heated water expands into it and pressure and level rise; the model holds pressure constant.
Transport time: one lap of 250 tonnes at 21,000 kg/s takes about 12 seconds, close to τ, a sign the model is rough on a scale of seconds.
Metal heat capacity: the vessel and piping store heat too, so real changes are slower.

Key takeaways

A simple model is not a wrong model; it only has to fit its question. Experts often sanity-check complex results with hand calculations like this. Temperature is changed by the difference between heat in and heat out, its speed is set by M·cp, and a model is trustworthy only when you know what it leaves out.