Liquid Cooling Loop for a Battery Pack: Pump, Radiator and Cold Plate

The problem

Car and bus packs and many high-power two-wheelers use liquid cooling because air cannot remove heat fast enough during fast charging or sustained high power. Engineers must choose coolant flow, cold-plate design and radiator size together, trading cell temperature against pump power and pressure drop. A system-level fluid–thermal model is the standard way to make those choices early.

Objectives

• Model a thermal-liquid cooling loop with pump, pipes, cold plate and radiator.
• Couple the loop to a lumped thermal model of the pack modules.
• Predict cell and coolant temperatures during a drive cycle and a fast-charge event.
• Size coolant flow rate and radiator capacity against pump power and pressure drop.
• Add a simple pump and fan control and compare energy use with constant operation.

What you need

Key concepts

Thermal liquid domain

Simscape's thermal-liquid domain tracks pressure, mass flow and temperature together, so pressure drop, pump work and heat transport are solved consistently.

Cold plate

An aluminium plate with internal channels under the cells. Heat flows from cell to plate by conduction through a thermal interface material and from plate to coolant by convection in the channels.

Coolant heat balance

Q = ṁ·cp·(T_out − T_in). For a given heat load, higher flow means a smaller coolant temperature rise but more pump power.

Radiator effectiveness

The ratio of actual heat rejected to the maximum possible, set by the radiator's size and the air and coolant flow rates. It limits how close coolant can get to ambient temperature.

Step-by-step guide

1. 1

   Set requirements and example parameters

   Take an example 30 kWh, 350 V car pack in eight modules, a 50/50 water–glycol coolant and a target of cells below 40 °C during a 1C fast charge at 40 °C ambient, with module-to-module spread below 5 °C. Estimate heat loads from I²R using example cell resistance. Write a parameter script with coolant properties, cold-plate area, channel hydraulic diameter, interface thermal resistance, pump flow range (for example 4–15 L/min) and radiator size.

   Checkpoint: A requirements table and parameter script with every value labelled as example or sourced.

2. 2

   Build the pack thermal side

   Model each module as a Thermal Mass with a heat source driven by a heat-load timeseries. Connect each module to its section of cold plate through a Thermal Resistance block representing the interface material and plate conduction. Keep the modules separate so you can see how coolant warming along the plate changes the spread. Test the thermal side alone with a fixed-temperature boundary in place of coolant.

   Checkpoint: With a fixed 25 °C boundary and a constant heat load, module temperatures settle at 25 °C plus Q·R_th as calculated by hand.

3. 3

   Build the coolant loop

   Set up the thermal-liquid network with a Thermal Liquid Settings (TL) block holding water–glycol properties, a pump from Simscape Fluids driven by a flow or speed command, Pipe (TL) segments for each cold-plate channel section, connecting hoses with realistic lengths and a reservoir. Connect each cold-plate pipe's thermal port to its module interface. Use a variable-step stiff solver such as ode23t and check the Simscape diagnostics for initialisation warnings.

   Checkpoint: With no heat load, the loop runs at the commanded flow with a pressure drop in a plausible range for the hose and channel lengths, and temperatures stay uniform.

4. 4

   Add the radiator and fan

   Add a radiator using a Simscape Fluids heat-exchanger block for thermal liquid, or a simplified effectiveness model built from a Pipe (TL) section with a convective boundary to ambient air. Make air flow depend on vehicle speed plus a fan contribution. Parameterise it from an example radiator size and check rejected heat at a known operating point against Q = ε·C_min·(T_coolant,in − T_air,in).

   Checkpoint: Radiator heat rejection at a test point matches the effectiveness calculation within about 10 per cent.

5. 5

   Verify the energy balance

   Run a steady case with a constant heat load and fixed flow. Check that heat into the coolant from the modules equals ṁ·cp·ΔT across the cold plate and equals heat rejected by the radiator once temperatures settle. Log heat flows and temperatures with the Simulation Data Inspector. An imbalance usually means a wrong boundary condition or that the run has not reached steady state.

   Checkpoint: Heat in and heat out agree within two per cent at steady state.

6. 6

   Simulate drive and fast-charge cases

   Run a highway drive cycle and a 1C fast-charge event (for example 40 minutes) at 30 and 40 °C ambient. For charging, set vehicle speed to zero so only the fan cools the radiator. Script the runs with sim() and collect the hottest module temperature, spread, coolant inlet and outlet temperatures and pump hydraulic power. Start each run from a stated initial temperature, for example pack and coolant at ambient after a soak, so cases are comparable.

   Checkpoint: Temperature traces that rise during charging and settle when the radiator capacity matches the heat load, with the worst case clearly identified.

7. 7

   Sweep flow rate and radiator size

   Sweep pump flow and radiator size over a grid for the worst case. Plot peak cell temperature, spread and pump power against flow. Identify the knee point beyond which extra flow gives little cooling benefit but rising pump power, and the minimum radiator size that meets the requirement at 40 °C ambient. Note that at high ambient the coolant cannot go below ambient without a chiller.

   Checkpoint: Contour or line plots of peak temperature and pump power with the chosen design point marked.

8. 8

   Add simple control

   Build a controller in Simulink that sets pump flow and fan speed from the hottest module temperature, for example off below 30 °C and rising linearly to maximum at 38 °C, with hysteresis. Compare peak temperature and total pump-plus-fan energy against constant full-flow operation over the drive cycle and charge event. Rate-limit the pump command so flow does not jump, and check that the controller does not cycle on and off repeatedly near its thresholds.

   Checkpoint: The controlled case meets the temperature requirement with lower auxiliary energy than constant operation.

9. 9

   Report results and limitations

   Document the loop architecture, parameters and sources, energy-balance verification, sweep results and controller comparison. State the limitations: one-dimensional pipe models, assumed convection coefficients in the channels, no chiller, and a lumped module representation. Recommend the flow rate and radiator size with the margin they leave at 40 °C ambient. Include the energy-balance checks as evidence that the model is trustworthy, and attach the scripts used for every sweep so the design point can be regenerated.

   Checkpoint: A report with a justified design point and a clear statement of what would need test data to confirm.

Testing & validation

• Fixed-boundary module test against hand calculation.
• Steady-state energy balance across cold plate and radiator.
• Radiator heat rejection against an effectiveness calculation.
• Peak temperature, spread and pump power for drive and fast-charge cases.
• Controlled versus constant operation energy comparison.

Expected results

A Simscape Fluids cooling-loop model coupled to the pack, a recommended pump flow and radiator size for the example pack, and a simple controller that meets the temperature target with less auxiliary energy.

Take it further

• Add a chiller loop from the air-conditioning circuit for very hot ambient.
• Compare series and parallel cold-plate channel layouts for temperature spread.
• Couple to an electrical pack model so heat generation responds to temperature.
• Add a PTC heater for cold-climate warm-up before charging.

Viva and review questions

Learn it at DIYguru

These DIYguru courses and programmes teach what this project needs.