The pursuit of efficient, cost-effective, and quiet power transmission solutions has led to the increased adoption of screw gears, particularly those employing a polymer-steel material pairing. In such configurations, the worm is typically manufactured from steel, while the worm wheel is injection-molded from engineering polymers such as Polyoxymethylene (POM). This combination offers significant advantages, including reduced weight, lower acoustic emissions, excellent electrical insulation properties, and the potential for high single-stage transmission ratios. Furthermore, the mass production of polymer worm wheels via molding presents substantial cost savings compared to traditional gear-generating machining methods for metal gears. However, the successful design and application of these screw gears hinge on a thorough understanding of their thermal behavior. The thermophysical and tribological properties of polymers, such as the elastic modulus and coefficient of friction, exhibit a much stronger dependence on temperature than those of metals. Consequently, accurate thermal analysis is paramount for determining the load-carrying capacity, ensuring operational temperatures remain below the polymer’s heat deflection or melting point to prevent thermal failure, and optimizing the gear design for compactness and extended service life.
Traditional thermal assessment methods for gears often involve empirical testing or localized measurements. While valuable, these approaches may not capture the holistic heat dissipation pathways within a gearbox assembly. The Thermal Network Method (TNM), based on the analogy between electrical and thermal circuits, provides a powerful framework for modeling complex thermal systems. It allows for the analysis of heat flow paths, the identification of critical thermal resistances, and the prediction of steady-state temperature distributions at defined nodes within the system. In this study, I employ a TNM approach implemented within the Simscape physical modeling environment to conduct a detailed steady-state thermal analysis of a polymer-steel screw gear pair. The analysis focuses on a gear set operating under natural air convection cooling, a common scenario in cost-sensitive applications like automotive auxiliaries and domestic appliances. The primary objectives are to model the complete heat transfer network, quantify the major sources of power loss, calculate the associated thermal resistances, and simulate the resulting temperature field and heat flow distribution to identify thermal bottlenecks and potential design improvements for these screw gears.
Fundamentals of Thermal Network Analysis for Screw Gears
The thermal network analysis of a mechanical system involves discretizing it into a finite number of isothermal nodes connected by thermal resistances, which represent the paths for conductive, convective, or radiative heat transfer. For the analyzed screw gear system, radiation is neglected due to the relatively moderate operating temperatures expected. The governing principle is analogous to Ohm’s law: a temperature difference (thermal potential, analogous to voltage) drives a heat flow rate (analogous to current) through a thermal resistance. The steady-state temperature at each node is determined by the balance of heat flows entering and leaving it.
I utilize MATLAB Simscape, an extension of Simulink for modeling physical systems, to construct and solve this thermal network. Simscape’s thermal domain library provides fundamental blocks such as ideal heat flow sources (for power losses), thermal resistances, thermal masses, and temperature sources (for boundary conditions like ambient temperature). Sensors for temperature and heat flow can be directly integrated into the model. This graphical modeling approach offers significant advantages over solving a system of algebraic equations manually. It provides an intuitive, visual representation of the thermal circuit, making it easier to modify the gear arrangement, adjust parameters, and directly observe the temperature at all nodes and heat flow through all branches simultaneously during simulation. The core principle of the Simscape-based thermal analysis for the screw gears is illustrated in the conceptual diagram below, where power losses are injected as heat flows, which then dissipate through a network of resistances to the ambient sink.
Geometric Configuration and Thermal Network Node Distribution
The subject of this analysis is a polymer-steel screw gear set from a test rig. The worm is made of 45# steel, and the worm wheel is made of POM. The transmission employs deep-groove ball bearings with double-sided seals (labeled A, B for the worm shaft and C, D for the wheel shaft). Key geometric and operational parameters are summarized in Table 1.
| Parameter | Symbol | Value | Unit |
|---|---|---|---|
| Transmission Ratio | $i$ | 40 | – |
| Wheel Helix Angle | $\beta_2$ | 7.5 | ° |
| Input Speed (Worm) | $n_1$ | 1500 | rpm |
| Center Distance | $a$ | 30 | mm |
| Worm Material | – | 45# Steel | – |
| Wheel Material | – | POM | – |
| Cooling Condition | – | Natural Air Convection | – |
Based on the geometry and expected heat flow paths, the system is discretized into 12 primary isothermal nodes, as shown in the node distribution schematic. The nodes are defined as follows: Node 0 (Ambient), Node 1 (Worm meshing zone), Node 2 (Wheel meshing zone), Nodes 3 & 4 (Midpoints of non-meshing sections of the worm shaft between the mesh and bearings A & B, respectively), Node 5 (Center of the worm wheel body), Node 6 (Interface between worm wheel and its shaft), Nodes 7-10 (Inner rings of bearings A, B, C, D), and Node 11 (Gearbox housing and bearing outer rings, assumed as one isothermal body due to the housing’s high aluminum conductivity).
The corresponding thermal network diagram is constructed, mapping all significant conductive and convective paths. The major heat sources are the frictional power loss from the screw gear mesh ($\dot{Q}_{m,worm}$, $\dot{Q}_{m,wheel}$) and the power losses from the four bearings ($\dot{Q}_{b,A}$, $\dot{Q}_{b,B}$, $\dot{Q}_{b,C}$, $\dot{Q}_{b,D}$). The thermal resistances ($R_{th,ij}$) connect these nodes, representing the impediment to heat flow between them.
Modeling of Power Losses in Screw Gears
Accurate quantification of heat generation is critical for thermal network analysis. For polymer-steel screw gears, the total frictional power loss in the mesh ($P_v$) is a dominant source. An empirical model derived from experimental studies for POM-steel pairs is adopted. The model accounts for the significant influence of sliding velocity and temperature on the friction coefficient. The total mesh loss is given by:
$$ P_v = \frac{2 \pi T_2 n_1}{30} \left( \frac{1}{\eta} – 1 \right) $$
where $T_2$ is the output torque (N·m), $n_1$ is the input (worm) speed (rpm), and $\eta$ is the transmission efficiency. The efficiency itself is a function of the gear geometry and a loss factor $\phi_{PZ}$:
$$ \eta = \frac{\tan(\beta_2)}{\tan(\beta_2) + \arctan(\phi_{PZ})} $$
The loss factor $\phi_{PZ}$ is empirically correlated with the sliding velocity $v_g$ (m/s) and the center distance $a$ (mm) through a complex relationship:
$$ \phi_{PZ} = \frac{ \nu_0 \sin^2(\beta_2) + \nu_1 \sin(\beta_2)^2 + \nu_2 v_g^2 + \nu_3 v_g^4 + \nu_4 a^2 + \nu_5 a^4 + \nu_6 v_g^2 a^2}{ \nu_7 v_g^2 a^4 + \nu_8 a^2 v_g^4 + \nu_9 } $$
The coefficients $\nu_0$ to $\nu_9$ for POM are listed in Table 2. This power loss $P_v$ is partitioned between the worm and wheel nodes (1 and 2) in the thermal model. The heat generation rate is $\dot{Q}_m = P_v$.
| $\nu_0$ | $\nu_1$ | $\nu_2$ | $\nu_3$ | $\nu_4$ |
|---|---|---|---|---|
| 10.864 | 2.8090E-3 | 0.97017 | 3.1811E-4 | 0.74456 |
| $\nu_5$ | $\nu_6$ | $\nu_7$ | $\nu_8$ | $\nu_9$ |
| 1.5181 | 6.6791E-4 | 0.036533 | 2.5208 | 0.26169 |
Bearing power losses are calculated using the well-established Palmgren method. The frictional torque $M$ (N·mm) for a bearing is:
$$ M = M_0 + M_1 $$
where $M_0$ is the torque due to viscous drag of the lubricant, and $M_1$ is the torque due to applied load. For a bearing with mean diameter $d_m$ (mm) and lubricant kinematic viscosity $\nu$ (mm²/s):
$$ M_0 = \begin{cases}
10^{-7} f_0 (\nu n)^{2/3} d_m^3 & \text{if } \nu n \geq 2000 \\
1.6 \times 10^{-7} f_0 d_m^3 & \text{if } \nu n < 2000
\end{cases} $$
$$ M_1 = f_1 P_1 d_m $$
Here, $f_0$ and $f_1$ are factors depending on bearing type and lubrication, $n$ is the bearing speed (rpm), and $P_1$ is the dynamic equivalent load (N). The total power loss for a bearing is $P_b = M \cdot n \cdot \pi / 30 \times 10^{-3}$ (W). In the thermal network, this loss is typically split equally between the inner ring (source at node 7-10) and the outer ring (source at node 11), as the heat generation is distributed at the rolling contacts.
Thermal Resistance Calculation Model
The thermal resistances connecting the nodes are calculated based on geometry, material properties, and heat transfer modes. The primary modes are conduction through solids and convection to the air.
1. Conduction Resistance: For a simple prismatic shape with cross-sectional area $A_c$, length $L$, and thermal conductivity $k$, the resistance is:
$$ R_{th,cond} = \frac{L}{k A_c} $$
This formula is applied to calculate resistances like $R_{th,13}$ and $R_{th,14}$ along the worm shaft, and $R_{th,25}$ through the worm wheel body.
2. Convection Resistance: For heat transfer from a surface at temperature $T_s$ to a fluid at $T_\infty$, the resistance is:
$$ R_{th,conv} = \frac{1}{h A_s} $$
where $h$ is the convective heat transfer coefficient and $A_s$ is the surface area. For natural air convection inside the gearbox, the surfaces of the worm, wheel, and shafts are modeled. The coefficient $h$ is often estimated using correlations for relevant geometries (e.g., horizontal cylinders for shafts). For the forced convection effect caused by the rotating components, a correlation for flow over a flat plate can be adapted, providing a Nusselt number $Nu$:
$$ Nu = \frac{h L_c}{k_{air}} = 0.644 Re^{1/2} Pr^{1/3}, \quad \text{for } Re < 10^5 $$
where $L_c$ is a characteristic length, $Re$ is the Reynolds number based on peripheral speed, $Pr$ is the Prandtl number of air, and $k_{air}$ is its thermal conductivity.
3. Bearing Interface Resistance: The thermal resistance between the bearing inner ring and the shaft ($R_{th,73}$, etc.) and between the outer ring and the housing ($R_{th,11,amb}$ via housing) involves contact conduction. These are often the most uncertain parameters and can be estimated based on fits and surface roughness or treated as calibration parameters.
4. Housing to Ambient Resistance: This is a combined natural convection and radiation resistance from the external surface of the gearbox to the environment. For simplification in this model focused on internal distribution, it is often represented by an overall effective resistance $R_{th,11,0}$.
A summary of key thermal resistances and their calculation basis is provided in Table 3.
| Thermal Resistance | Description | Primary Mode | Governing Parameters |
|---|---|---|---|
| $R_{th,12}$ | Worm-Wheel contact interface | Conduction | Contact area, interface material properties. |
| $R_{th,13}$, $R_{th,14}$ | Along worm shaft (mesh to bearing) | Conduction | Shaft length, diameter, steel conductivity. |
| $R_{th,25}$ | Through worm wheel body | Conduction | Wheel geometry, POM conductivity. |
| $R_{th,1,air}$, $R_{th,2,air}$ | From gear surfaces to internal air | Convection | Surface area, local air speed (rotation-induced). |
| $R_{th,73}$, $R_{th,84}$ | Bearing inner ring to worm shaft | Contact Conduction | Interference fit, surface finish. |
| $R_{th,11,0}$ | Gearbox housing to ambient | Combined Convection/Radiation | Housing external area, ambient conditions. |
Simulation Analysis and Results
Implementing the complete thermal network model with the calculated power losses and thermal resistances in Simscape allows for system-level simulation. The steady-state temperatures at all nodes are solved simultaneously by the software’s solver. I analyze the results for varying output torque $T_2$, which directly scales the primary heat generation from the screw gear mesh.
Temperature Distribution: The simulation results reveal a distinct and consistent thermal pattern. The temperature at the worm meshing zone (Node 1) is significantly higher than at any other node within the system, including the polymer worm wheel meshing zone (Node 2). This is a direct consequence of the vastly different thermal conductivities of steel (≈50 W/m·K) and POM (≈0.3 W/m·K). The frictional heat generated at the interface cannot easily conduct into the polymer wheel; instead, the majority of it flows into the highly conductive steel worm. Consequently, the worm acts as the primary heat sink and thermal pathway from the mesh. This creates a substantial temperature gradient across the contact interface, meaning the limiting factor for preventing thermal failure of the POM wheel is the interface temperature, which is closely approximated by the worm temperature (Node 1).
The temperatures of all nodes increase approximately linearly with increasing output torque, as shown in the simulated trend. The slope is steepest for the worm temperature, indicating its high sensitivity to load. The bearing inner ring temperatures (Nodes 7-10) are notably lower than the worm temperature, highlighting the effectiveness of the worm shaft as a heat conduction path away from the hot spot.
Heat Flow Analysis: Probing the heat flows in the Simscape model quantifies the dissipation pathways. The analysis confirms that over 90% of the heat generated at the screw gear mesh flows into the worm (Node 1). From there, the dominant dissipation path is via conduction along the worm shaft to the nearby bearings (Nodes 3 & 4, and subsequently to Nodes 7 & 8), and from the bearings to the housing (Node 11). Convective heat transfer from the worm surface to the internal air accounts for a minor portion. In contrast, heat flow into the polymer wheel is minimal, and its subsequent dissipation is negligible in comparison. This stark imbalance underscores that for this type of screw gear, thermal management is essentially “worm-centric.” Improving the overall thermal performance requires focusing on enhancing the heat extraction from the worm, not the wheel.
The large temperature drop along the worm shaft (from Node 1 to Nodes 3/4) signifies that the conductive resistance along the shaft plays a major role. This points to two straightforward design strategies for thermal improvement: 1) Using a worm material with higher thermal conductivity, and 2) Reducing the conduction path length by minimizing the distance from the mesh to the supporting bearings.
Parametric Study and Design Implications
To validate the insights from the heat flow analysis, I conducted parametric studies by modifying the Simscape model. Two specific changes were simulated: first, increasing the thermal conductivity of the worm material by 5 W/m·K (simulating a higher-grade alloy), and second, reducing the length of the worm shaft (distance from mesh to bearing) by 2 mm.
The results are illuminating. At a high operational torque, the model with the higher-conductivity worm material predicted a steady-state worm temperature approximately 10 K lower than the baseline model. Similarly, the model with the shortened worm shaft predicted a temperature reduction of about 5 K under the same high load. These improvements directly translate to a higher permissible load for a given maximum operating temperature or allow for the use of a smaller gear set to achieve the same load capacity, leading to cost and space savings.
Furthermore, the material properties of the polymer, especially its thermal conductivity and specific heat capacity, also influence the results, though to a lesser extent than the worm’s properties. The convective heat transfer inside the gearbox, dependent on the arrangement and rotational speeds of the components, is another area for potential optimization, though its impact is secondary to conduction through the worm.
The thermal network model, therefore, serves as a powerful tool for the design engineer working with polymer-steel screw gears. It moves beyond guesswork, enabling quantitative comparisons of different material choices, geometric layouts, and even the potential benefit of adding simple cooling fins to the worm shaft or housing. The primary recommendation stemming from this analysis is that thermal design efforts should prioritize minimizing the thermal resistance from the worm meshing zone to the gearbox housing. Selecting a high-conductivity steel for the worm, minimizing the worm shaft span between bearings, and ensuring good thermal contact at the bearing-housing interfaces are the most effective strategies.
Conclusion
In this study, I have developed and demonstrated a comprehensive thermal network modeling methodology for analyzing the steady-state heat transfer in polymer-steel screw gears. Utilizing the Simscape physical modeling environment provided an efficient and intuitive platform to construct, simulate, and analyze the complex thermal circuit representing the gearbox assembly. The model integrates empirical formulas for the power losses specific to POM-steel screw gear meshes and standard calculations for bearing losses and thermal resistances.
The simulation results unequivocally identify the steel worm as the critical thermal component. Due to its high thermal conductivity relative to the polymer wheel, it absorbs the vast majority of the frictional heat generated at the mesh. The subsequent heat dissipation occurs predominantly via conduction along the worm shaft to the bearings and housing, rather than through convection to the internal air or into the wheel. This makes the conductive thermal resistance along the worm shaft a primary bottleneck.
Consequently, the thermal performance and load-carrying capacity of such screw gears can be most effectively enhanced by focusing on the worm’s thermal design. The parametric studies confirm that employing a worm material with higher thermal conductivity and/or reducing the length of the worm shaft between the mesh and its support bearings are two practical and impactful design modifications. These changes directly lower the maximum operating temperature at the critical mesh interface, thereby either allowing for higher load transmission or enabling a more compact gear design for a given load. This thermal-network-based approach provides a valuable analytical foundation for optimizing the design and application of cost-effective and reliable polymer-steel screw gears across various industries.