Stirling Engine
Stirling Engine
A gamma Stirling engine is a closed-cycle, external-combustion heat engine where a displacer piston shuttles gas between a hot and cold region, while a separate power piston converts the resulting pressure changes into mechanical work.
Stirling engines in general (including alpha and beta types) are used in more demanding applications such as solar dish power systems, industrial waste-heat recovery, silent generators, submarine propulsion, cryogenic coolers, and residential combined heat-and-power units. Overall, gamma engines are ideal for demonstration and small-scale mechanical work, while more advanced Stirling designs serve specialized industrial, scientific, and energy-generation roles.
The Stirling engine cycle is a repeating process where a sealed amount of gas is heated and cooled to make the engine run. First, the gas is moved to the hot side, where it heats up, expands, and pushes the power piston outward to produce work. Next, the gas is shifted to the cold side, where it cools down and its pressure drops. The piston then moves inward easily, compressing the cool gas with little effort. Finally, the gas is moved back toward the hot side, warming up again and preparing for the next expansion.
This continuous sequence of heating → expansion → cooling → compression creates a loop that drives the engine smoothly, using any external heat source.
My goal was to build a gamma Stirling engine that successfully demonstrated the Stirling cycle. I designed and assembled the displacer cylinder, power piston, crankshaft, and regenerator, ensuring the correct phase difference so the gas moved between hot and cold sides properly. By minimizing dead volume and optimizing heat transfer, the engine ran smoothly, clearly showing the heating, expansion, cooling, and compression steps. Completing the project gave us hands-on experience in thermodynamics, and mechanical design, and it was rewarding to see outher engine operate successfully.
The drive train of a Stirling engine MUST
Maintain continuous piston motion
Ensure proper 90-degree phase relationship between pistons
Minimize mechanical energy loses (i.e. friction)
Flywheel and Drivetrain Design
The flywheel is a critical component for maintaining stable and continuous operation in the Stirling engine. Real engines deviate from ideal Stirling cycle assumptions due to finite heat transfer, thermal resistances, and temperature gradients. These effects produce fluctuations in gas pressure and piston force, which can lead to intermittent or unstable motion. The flywheel provides rotational inertia, storing energy during high-torque periods and releasing it during low-torque portions of the cycle, thereby smoothing torque fluctuations and preventing stalling.
Both the power and displacer pistons are mechanically coupled to a common drive shaft through the flywheel. The flywheel’s inertia maintains angular momentum through regions of low or negative torque, enabling continuous piston motion despite variations in thermodynamic forcing. This inertial coupling is essential for overcoming transient losses and ensuring consistent engine operation.
Shaft and Bearings
The drivetrain consists of a single shaft supported by two ball bearings mounted in an aluminum frame. The flywheel is rigidly attached to the drive shaft using clamping flanges, ensuring that the shaft and flywheel rotate as a single rigid body. Spacers were designed to seat against the inner races of the bearings, preventing excessive bearing preload and minimized frictional losses. The use of 3D-printed spacers also allowed precise control of spacer length for maintaining shaft alignment and preventing unintended side loads that could introduce parasitic forces into the linkage system.
Piston Phase Control and Linkages
Proper Stirling engine operation requires a 90-degree phase offset between the power and displacer pistons. To maintain this relationship:
A custom clamp indexes both piston linkages to the drive shaft at predetermined angular positions
Each linkage uses a 4-40 screw as a crank pin
Ball bearings and 3D-printed spacers are incorporated to reduce friction and allow smooth relative motion
Nylon lock nuts prevent loosening due to vibration
This arrangement ensures that the pistons move through their full stroke range without interference and maintains consistent kinematics.
Flywheel Geometry and Material
Diameter: 3.0 in
Material: Aluminum, chosen for low cost, manufacturability, and adequate stiffness
The flywheel design incorporates modular stainless-steel perimeter weights:
Attachment method: Screws and nylon lock nuts
Purpose: Adjust total rotational inertia without redesigning the flywheel
One important consideration before discussing the overall design rationale of the system is a key specification that was known to the me early in the project. According to the manufacturer’s specification sheet, the selected power piston is capable of producing a maximum force of approximately 30 lbf before failure. Consequently, a large portion of the drivetrain and supporting components were designed and specified around this limiting force to ensure safe and reliable operation.
Flywheel Inertia
The flywheel was fabricated from aluminum so that it could be readily waterjet cut while still providing sufficient stiffness and dimensional stability for my application. Using SolidWorks, the area moment of inertia of the original flywheel geometry was calculated and found to be approximately 4.28 Kg*m^2. Increasing the rotational inertia of the flywheel would improve rotational smoothness and assist with engine startup by reducing cyclic speed fluctuations.
To increase the flywheel inertia while maintaining design flexibility, 1/8-inch steel plates were sourced and water-jet cut into removable weights that were fastened to the flywheel. SolidWorks analysis showed that each added weight increased the flywheel’s rotational inertia by approximately 11.01 Kg*m^2. Given enough space to add 4 sets of weights, this totals to a ~93% increase in rotational inertia (totaling 48.6 Kg*m^2) This modular approach allowed the total inertia to be adjusted experimentally, preventing the system from being over-initialized in the event that the engine output torque was lower than predicted. This precaution was especially important given that preliminary empirical testing of a smaller demonstration engine indicated that first-order thermodynamic models tended to overestimate actual performance.
Bearing Selection
Bearing selection was a critical aspect of the mechanical design, as the bearings directly influence frictional losses and overall system reliability. Each bearing was expected to experience a radial load on the order of the piston force, approximately 30 lb. Bearings with a dynamic radial load capacity of approximately 140 lbs were selected. This resulted in a safety factor of approximately 4.7, providing sufficient margin for continuous operation and accounting for uncertainties in loading, alignment, and transient forces. Open bearings were selected so they could be lubricated if needed.
SolidWorks Motion Analysis
Linkage dimensions were based on reference designs from similar Stirling engines and fabricated from aluminum to minimize mass while providing adequate strength.
Kinematic compatibility was the primary design concern. A motion study in SolidWorks defined the geometry of both the power and displacer linkages, ensuring that each piston achieved its required stroke and maintained the correct phase relationship. Linkage lengths were adjusted so that maximum piston travel was limited to 1/8 in short of the cylinder ends, preventing interference and mechanical damage. These constraints established the final linkage geometry and ensured reliable operation throughout the full range of motion.
Sizing the Displacer Chamber/Piston
I derived the displacer piston dimensions using standard Stirling engine ratios which I validated through research and a MATLAB based simulation.
Power Piston Reference Dimensions
Stroke Length - 1.4in
Piston Area - 0.4in^2
Displacer Piston Dimensions
Power Piston Reference Dimensions
Stroke Length - 1.4in
Piston Area - 0.4in^2
Displacer Piston OD - 1.75"
This dimension has huge impacts on the pistons swept volume and chamber size.
The displacer piston has a swept volume three to six times larger than the power piston because its job is to move most of the working gas between the hot and cold spaces, not to produce power. The larger volume ensures effective gas transfer and strong pressure changes, improving efficiency. The power piston only needs a smaller swept volume to convert these pressure changes into useful work, since a larger one would increase losses without added benefit.
Displacer piston swept volume ≈ 3 to 6 times the power piston swept volume.
Swept Volume Equation = Piston Area * Stroke Length
Displacer Swept Volume Max and Min
Swept_Min = 0.389in^3 * 3 = 1.167in^3
Swept_Max = 0.389^3 * 6 = 2.335in^3
Displacer Piston OD Min = 1.364in
Displacer Piston OD Max = 1.842in
Displacer Piston Stroke Length - 0.4" (Equal to Power Piston)
Dependent on the Power Piston
Displacer Piston Thickness - 1/16"
Keeps Piston Lightweight
Displacer Piston Chamber Length ~ 120mm
The chamber must accommodate the piston along its entire cycle, with enough space for internal components and air to travel around it without being compressed. Too much space results in unused working air, and too little space will compress the air in the chamber without directing it to the power piston once again losing a lot of mechanical output.
Displacer Piston Length + (Displacer Piston Stroke Length * 1.1) ~ 120mm ~ 4.75"
The 10% clearance on either end of the piston keeps space for the bearing mount and air circulation
Displacer Piston Chamber ID - 2"
The chamber diameter must have a roughly 10% clearance to allow the working air to travel between the piston zones
Displacer Piston OD * 1.1 ~ 1.925", thus I selected 2" pipes on Mcmaster
Low Friction
To satisfy the low-friction requirements of the displacer piston I used a high-temp dry-sleeve bearing This bearing is used to guide the piston shaft which then connects to the engine drivetrain.
Since the displacer piston does not require an air-tight seal I did not have to consider friction of an o-ring or any other mechanical energy losses.
Crankshaft Phase Shift
Why is the 90 degree phase angle ideal?
The displacer piston must move the gas before the power piston reacts.
The 90 degree phase angle between pistons is described within the drivetrain subsystem.
Displacer Piston Conduction
How did I ensure the displacer piston does not reduce the temperature gradient of the hot and cold regions?
By making the displacer piston hollow, I avoided conducting heat between the chamber ends. The air inside of the piston acts as an insulator, stopping the piston from reaching either temperature extreme.
Minimizing Dead Volume
How did I make the engine more efficient?
Minimizing dead volume in a Stirling engine is important because unused space reduces pressure changes in the gas, lowering power and efficiency. Smaller dead volume ensures more gas participates in expansion and compression, producing stronger piston motion, smoother operation, and better demonstration of the Stirling cycle.
To validate my research and predict key system behaviors such as internal pressure, power output, and rotational speed, I used a pre-made MATLAB Simulink model of a Stirling engine. Because the model operates on nearly identical principles to the design, it could be adapted by modifying the governing equations along with most parameters and dimensions.
However, the model became unstable when full-scale engine dimensions were applied. To continue using the tool, I reverted to smaller dimensions and systematically adjusted selected parameters to study their effects on system performance. This parametric analysis allowed us to understand how factors such as flywheel mass, dead volume, and displacer-to-power piston sizing influence engine behavior. The insights gained helped us prioritize critical design features, particularly the optimal displacer piston ratio relative to the power piston.
Finally, I scaled the simulated internal pressure results from the smaller model to estimate pressures for the full-size engine, which informed the design of the pressure vessel and the selection of tubing and push-to-fit connectors.
The Power Piston of a Stirling engine MUST
Have an airtight seal
Have low friction
Be strong and durable enough to withstand high rpms
Be thermally stable
Have minimal dead volume
Have a 90 degree phase angle from the displacer piston
Since the power piston was a COTS (component off the shelf), I did not have to be concerned about a majority of power piston requirements excluding the 90 degree phase angle between each piston. This was a big relief to the budget and gave me some constraints for the rest of the system.
Piston Dimensions
As seen at right, having a fixed power piston meant that I needed to spec the system around the piston's dimensions and geometry. The frame needed to include a space to hold the piston, the tubing size came from air port on the piston, and the displacer geometry needed to displace enough air (and not much more than) to push the power piston to a full stroke. Similarly to this, the drivetrain system had to accommodate the stroke length of the power piston and match that stroke length to the displacer stroke (offset in phase by 90 degrees). Finally, because I had the threaded ball joint end configuration, I needed to drill and tap a hole perpendicular to the drivetrain shaft axis.
Piston Dimensions
As seen at right, having a fixed power piston meant that I needed to spec the system around the piston's dimensions and geometry. The frame needed to include a space to hold the piston, the tubing size came from air port on the piston, and the displacer geometry needed to displace enough air (and not much more than) to push the power piston to a full stroke. Similarly to this, the drivetrain system had to accommodate the stroke length of the power piston and match that stroke length to the displacer stroke (offset in phase by 90 degrees). Finally, because I had the threaded ball joint end configuration, I needed to drill and tap a hole perpendicular to the drivetrain shaft axis.
Confirmation of Component Selection
While it was extremely convenient that there was a power piston from the thermodynamics department, I did need to confirm that this piston would indeed work for the system. I had a 1" stroke length, and the piston itself is small, but is completely airtight and can pack a punch: it can output 30.9 pounds maximum, and it can take 100 psi. I was not going to reach this: atmospheric pressure is 14.7 psi at 288.15 degrees Kelvin. Using PV = nRT, we know that as temperature increases, (because the system will be airtight), the pressure will increase by the same factor. I can estimate with this model that I would have to get the air within the system to reach that 100 psi, I would need to get the air inside to (100/14.7)*288.15 = 1960.2 degrees Kelvin, or 3068 degrees Fahrenheit. Seeing as the maximum theoretical flame temperature of isobutane, the fuel source is only 3587°F, it will be safe. The aluminum will move away the heat faster than is necessary for the air inside to reach anywhere close to that temperature. I was not concerned with the vacuum specs, because when the piston draws back, the displacer is doing the force action on the drivetrain.
The Heatsink of a Stirling Engine MUST
Maintain the temperature gradient between the hot and cold ends of the displacer piston
Provide stable heat flux away from cold end
Low thermal resistance/Good thermal conductivity
High-surface-area fin array
Optional fan to improve cooling
Maintain Temperature Gradient (By Cooling One End of the Displacer Chamber)
I decided to use a fan to cool the displacer chamber, in combination with a high-surface-area fin array. This design choice supported my goals of a self-contained Stirling Engine which wouldn't need any coolant reservoirs.
Originally the heatsink was a component clamped to the displacer chamber. I realized that combining these components into one piece would increase the heat-rejecting capacity. This also meant I would not have to consider varying thermal expansion coefficients.
Fan Configuration
The two fans are arranged in series since this will increase the cooling efficiency of the system. The two fans static pressures add together, allowing them to push air through restrictions (heatsinks, filters, ducts) much better than a single fan. The downstream fan also receives a smoother, faster airflow from the upstream fan, reducing turbulence and preventing aerodynamic stall, which improves overall airflow under load even if free-air CFM changes little.
Fabrication and Implementation
The heatsink/displacer chamber was turned on a lathe into spec. This gave me great control over chamber OD and fan-array dimensions. The mounting holes were drilled on the CNC mill.
I fortunately did not have to worry too much about the chamber's ID, as the piston would not be physically interfacing with it.
I used a MATLAB based simulation to decide the initial heatsink dimensions. After getting a general understanding of successful heatsink dimensions I then performed simulations in COMSOL to finalize the design. I found the heat flux into the system using a research paper that explored heat flux from a gas flame on a pot and then used COMSOL's fluid flow capabilities to use the forced convection coefficient to provide a heat flux out of the system. This allowed me to predict the thermal behavior of the system, where I ended up with a simulated steady state 120 degrees Kelvin temperature gradient.
Then I plugged the final heatsink dimensions into MATLAB where I was met with a successful simulated Stirling Engine. This gave me the comfort necessary to move forward with finalizing the CAD assembly and fabricating the design.
The Combustion Subsystem MUST
Have flow control
Be easy to ignite
Have considerable distance between flame and isopro canister
Remain leak free
Be rechargeable
Be capable of heating the displacer up to at least 300c
Be self-contained within the engine frame
Earlier I decided that the focus of the project would be on the mechanical output of the Stirling Engine, not necessarily in the fabrication of individual components. This meant I was okay searching for COTS to fulfill a lot of design choices and requirements.
I settled on a camping stove burner to satisfy the flow control and ignition requirements. Additionally, I'd have a pretty good estimate of the amount of heat added to the system, allowing me to move forward analytically without needing to test and validate a custom burner design.
Pictured to the right is the final combustion subsystem. COTS are mounted to the frame leading to a high power, controllable, easy to ignite, self-contained heating element.
I encountered several compatibility issues between fuel sources and adapters, largely due to insufficient specificity in manufacturer documentation. Although Lindal valves, specifically EN417 valves, are often listed as compatible, they actually exist in multiple sizes that are difficult to distinguish visually on most vendor websites. As a result, components that appeared compatible based on published specifications were either too large or too small to mate properly. To resolve this issue, I sourced a hose assembly with a removable adapter, which allowed me to use a fitting appropriately sized for the 100 g ISOPRO canister.
Because the flame was intended to operate for extended periods, safety was a primary concern. ISOPRO vapors are heavier than air, meaning that even a small leak could allow fuel to accumulate beneath the engine and create a significant explosion hazard. To mitigate this risk, I leak tested every connection and potential failure point using soapy water at each stage of assembly, as well as before every operating session. In addition, I designed the engine frame to position the ISOPRO canister well away from the flame. This prevented heat conduction through the frame from warming the canister walls, which could otherwise weaken the container and increase the risk of rupture or unintended ignition.
During testing, I implemented strict safety precautions, with particular emphasis on adequate ventilation, appropriate personal protective equipment, thorough leak checks, and having fire extinguishers readily available in the event of an emergency. Together, these precautions allowed me to operate the engine safely for long duration tests while minimizing the risk of leaks, overheating, or combustion related accidents.
The Frame of a Stirling Engine MUST
Provide a rigid structural framework to support and accurately align all engine subsystems (pistons, crankshaft, bearings, and flywheel).
Tolerate/Dissipate heat (not acting as the primary heat transfer component).
Ensure adequate clearance for moving components.
Easy to Assemble, manufacture, and test.
Aluminum frame used to structurally support and align all engine subsystems.
Slotted mounting features incorporated to simplify assembly, reduce hardware, and allow displacer removal.
Frame geometry provides sufficient rigidity and clearance to maintain alignment during operation.
Layout enables easy access to components for troubleshooting, adjustment, and iterative assembly.
Flywheel mounted at the end of the frame, utilizing the overall engine length to promote stable rotation.
Material Selection
The primary role of the frame was to hold together the remaining subsystems rather than to act as a performance-critical component. Based on the combustion specifications, it was anticipated that the engine would experience significant localized heating from a large flame. For this reason, aluminum was selected for the frame despite its higher cost.
Using aluminum throughout the engine ensured similar coefficients of thermal expansion across components, reducing the risk of misalignment or tolerance changes during operation. This was especially important because slotted features were used in the frame to simplify assembly and reduce hardware; consistent thermal expansion behavior was required to maintain proper fit and alignment.
The frame was also designed to allow rapid access to components for troubleshooting and adjustment. This enabled quicker diagnosis of issues and iterative refinement during assembly and testing, improving overall development efficiency.