Physics of the Stirling Engine
A thesis on the power output of this wonderful mechanism
For my undergraduate physics thesis, I studied the power output of the Stirling engine.
The goal of my research was to determine the efficacy of the Senft model, the salient thermodynamic model for the Stirling engine, in predicting the actual power output of an alpha-type engine. To accomplish this, I designed and machined select engine components that would allow me to manipulate key engine parameters to independently determine their effects on torque and power curves, which I could then compare against the corresponding effects theorized by the model.
Click below to download a PDF of my full thesis.
Original engine
This engine was powered by an alcohol burner, visible to the far left. Engine architecture is non-adjustable throughout.
Modified engine
Multi-position flywheels enable manipulation of the compression ratio and engine timing, a fiberglass-insulated, nichrome wire heating element offers precise temperature control, and a reflective patch on the far flywheel facilitates laser RPM measurement.
Component breakdown
New flywheels, heating element, and connecting rods are to the left and bottom. Modified base plate with ceramic heater clamp is to the right. The ceramic can withstand the heat of the nichrome coil without causing an electrical short.
Test setup
The box fan (bottom) and nichrome heater maintain a large yet stable temperature differential across the engine. The black arm extending left off the engine is a modified de Prony brake, a simple dynamometer that measures torque output at a given engine RPM. When both torque and RPM are known, power output at that RPM can be determined computationally.
Typical torque and power curves
After feeding collected data through a custom Mathematica module, I produced the following torque vs. RPM (blue) and power vs. RPM (red) curves. Torque output falls off with increasing RPM due to reduced thermal expansion within the engine per unit time. Power output rises to a peak, then falls off with rising RPMs due to increasing frictional, thermal, and entropic losses.
Conclusions
- The Senft model cannot be used to predict power output as it does not account for frictional, thermal, and entropic losses at higher RPMs.
- It is also unreliable as it does not account for mechanical losses from the reciprocating pistons and connecting rods.
- The model does effectively predict broad trends in power output resulting from different engine configurations. For this reason, it can be used to optimize an engine for maximum power or torque output.