Gear transmission serves as a critical mechanism for transmitting speed and torque in mechanical systems, valued for its reliability, stability, and compact design. Within this domain, the cylindrical gear is a fundamental component. Compared to traditional machining methods like cutting, the through-feed rolling process presents a promising near-net-shape manufacturing technique for cylindrical gears, offering advantages such as high production efficiency, minimal material waste, and enhanced mechanical properties. The rolling die is the core component of this forming equipment. However, conventional dies often encounter significant challenges during the through-feed rolling of cylindrical gears, including difficulty in initial tooth tip penetration into the workpiece, inadequate material flow, and severe stress conditions. To address these issues, this work proposes a novel geometric design methodology for a beveloid rolling die.
The core innovation lies in designing a rolling die with a continuously varying tooth thickness along its axis. This is achieved by conceptually simulating the process of generating a cylindrical gear workpiece with a continuously changing profile shift coefficient using a rack tool. The three-dimensional tooth surface of the workpiece during axial feed is reconstructed by arranging these profiles along the axis. Subsequently, applying the principles of gear meshing under a constant center distance yields the complete tooth surface equations for the corresponding beveloid rolling die. The die is logically segmented into three functional zones: a cutting section, a finishing section, and an exiting section, each with distinct design objectives.
Geometric Design Principle for the Beveloid Rolling Die
The design is fundamentally based on the gear generation principle using a rack cutter. By systematically varying the center distance between a standard rack tool profile and the workpiece blank, a series of cylindrical gear profiles with different modification coefficients (x) are obtained. These profiles represent the instantaneous cross-section of the deforming workpiece at different stages as it passes through the rolling die. By spatially arranging these profiles along the axial direction and performing a three-dimensional reconstruction, the theoretical deformed surface of the cylindrical gear workpiece during through-feed rolling is obtained. Finally, considering that the cylindrical gear workpiece and the rolling die mesh at a constant center distance, the conjugate tooth surface of the beveloid rolling die is derived using gear meshing theory, resulting in the complete three-dimensional model.
Coordinate System Establishment
To mathematically describe the generation process, multiple coordinate systems are established. A fixed coordinate system $S_0(o_0 -x_0y_0 )$ is set at the pitch point of the rack and workpiece. A moving coordinate system $S_1(o_1 -x_1y_1 )$ is attached to the rack tool, translating along its pitch line. Coordinate system $S_2(o_2 -x_2y_2 )$ is fixed at the center of the cylindrical gear workpiece, while $S_3(o_3 -x_3y_3)$ rotates with it. Similarly, $S_4 ( o_4 – x_4y_4 )$ is fixed at the rolling die center, and $S_5(o_5 -x_5y_5)$ rotates with the die.
The transformation matrix from the rack coordinate system $S_1$ to the workpiece coordinate system $S_3$ is given by the product of successive transformations:
$$ M_{31} = M_{32} \cdot M_{20} \cdot M_{01} = \begin{bmatrix} \cos\varphi & -\sin\varphi & r_1\cos\varphi + T\sin\varphi \\ \sin\varphi & \cos\varphi & r_1\sin\varphi – T\cos\varphi \\ 0 & 0 & 1 \end{bmatrix} $$
where $\varphi$ is the rotation angle of the cylindrical gear workpiece, $r_1$ is its pitch radius, and $T$ is the linear displacement of the rack tool.
The transformation matrix from the rack coordinate system $S_1$ to the beveloid rolling die coordinate system $S_5$ is:
$$ M_{51} = M_{54} \cdot M_{42} \cdot M_{20} \cdot M_{01} = \begin{bmatrix} \cos\varepsilon & \sin\varepsilon & – r_2(\cos\varepsilon + \varepsilon\sin\varepsilon) \\ -\sin\varepsilon & \cos\varepsilon & r_2(\sin\varepsilon – \varepsilon\cos\varepsilon) \\ 0 & 0 & 1 \end{bmatrix} $$
where $\varepsilon$ is the rotation angle of the rolling die and $r_2$ is the corresponding pitch radius of the beveloid die section.
Mathematical Model for Die Tooth Profile
The standard rack tool profile consists of three segments: the tip straight line (ab), the tip fillet curve (bc), and the flank straight line (cd). The equations for these segments in the rack coordinate system $S_1$, incorporating the variable modification coefficient $X$, are defined as follows for the cutting and exiting sections (For the finishing section, $X=0$).
Tip Line (ab):
$$ R^1_{ab} = \begin{bmatrix} R^1_{ab_x} \\ R^1_{ab_y} \\ 1 \end{bmatrix} = \begin{bmatrix} \pm ( -\pi m – 4X m \tan\alpha + l_{1_{ab}}) \\ – h^*_a m – X m + r_{bc}\sin\alpha – r_{bc} \\ 1 \end{bmatrix} $$
Tip Fillet (bc):
$$ R^1_{bc} = \begin{bmatrix} R^1_{bc_x} \\ R^1_{bc_y} \\ 1 \end{bmatrix} = \begin{bmatrix} \pm \left( -\frac{\pi}{2}m + 2X m \tan\alpha – h^*_a m \tan\alpha – X m \tan\alpha – r_{bc}\cos\alpha + r_{bc}\sin\theta_{2_{bc}} \right) \\ – h^*_a m – X m + r_{bc}\sin\alpha – r_{bc}\cos\theta_{2_{bc}} \\ 1 \end{bmatrix} $$
Flank Line (cd):
$$ R^1_{cd} = \begin{bmatrix} R^1_{cd_x} \\ R^1_{cd_y} \\ 1 \end{bmatrix} = \begin{bmatrix} \pm \left( -\frac{\pi}{2}m + 2X m \tan\alpha – h^*_a m \tan\alpha – X m \tan\alpha + l_{cd}\sin\alpha \right) \\ l_{cd}\cos\alpha – h^*_a m – X m \\ 1 \end{bmatrix} $$
In these equations, $m$ is the module, $\alpha$ is the pressure angle, $h^*_a$ is the addendum coefficient, $r_{bc}$ is the fillet radius, $l_{1_{ab}}$, $l_{cd}$, and $\theta_{2_{bc}}$ are parameters defining location on the curve, and the $\pm$ sign corresponds to the left/right flank of the cylindrical gear.
By applying the coordinate transformation $M_{31}$ to the rack profile equations $R^1$, the generated tooth profile of the cylindrical gear workpiece $W^1$ in its rotating coordinate system $S_3$ is obtained through the meshing condition. This workpiece profile $W^1$ is then used as the conjugate target. Applying the inverse transformation derived from $M_{51}$ yields the final tooth profile equations for the beveloid rolling die $B^1$ in its coordinate system $S_5$. The general form for transforming a point from the rack to the die is:
$$ B^1 = \begin{bmatrix} B^1_x \\ B^1_y \\ 1 \end{bmatrix} = \begin{bmatrix} R^1_x \cos\varepsilon + R^1_y \sin\varepsilon – r_2(\cos\varepsilon + \varepsilon\sin\varepsilon) \\ -R^1_x \sin\varepsilon + R^1_y \cos\varepsilon + r_2(\sin\varepsilon – \varepsilon\cos\varepsilon) \\ 1 \end{bmatrix} $$
This process is applied separately to define the profiles for the three sections of the beveloid die:
- Cutting Section: The modification coefficient $X$ varies from a negative starting value to 0 along the axis, facilitating gradual material penetration.
- Finishing Section: The modification coefficient is held constant at $X=0$, ensuring full-depth, precise teeth on the cylindrical gear.
- Exiting Section: The modification coefficient varies from 0 to a slightly negative value, creating clearance to prevent scratching the formed cylindrical gear teeth during retraction.
Analysis of Beveloid Rolling Die Geometric Parameters
To validate the proposed design methodology and analyze the effect of the modification coefficient range, a specific case study is conducted. The geometric parameters for the example cylindrical gear and die are summarized below.
| Geometric Parameter | Value | Geometric Parameter | Value |
|---|---|---|---|
| Module, $m$ (mm) | 1.75 | Pressure Angle, $\alpha$ (°) | 20 |
| Number of Workpiece Teeth, $z_w$ | 46 | Addendum Coefficient, $h^*_a$ | 1 |
| Number of Die Teeth, $z_d$ | 128 | Dedendum Clearance Coefficient, $c^*$ | 0.25 |
| Cutting Section Face Width (mm) | 40 | Finishing Section Face Width (mm) | 25 |
| Exiting Section Face Width (mm) | 5 | Workpiece Initial Diameter (mm) | Φ80.75 |
Three distinct beveloid rolling die configurations, defined by their modification coefficient ranges, are designed and analyzed via Finite Element Method (FEM) simulation using DEFORM-3D™:
- Die I: Cutting: $X = -2 \to 0$, Finishing: $X = 0$, Exiting: $X = 0 \to -0.5$.
- Die II: Cutting: $X = -1 \to 1$, Finishing: $X = 1$, Exiting: $X = 1 \to 0.5$.
- Die III: Cutting: $X = 0 \to 2$, Finishing: $X = 2$, Exiting: $X = 2 \to 1.5$.
The FEM model considers the plastic deformation of an AL-6061-T6 workpiece, modeled with tetrahedral elements and local mesh refinement in the deformation zone. The rolling dies, mandrel, and restraint plates are modeled as rigid bodies. The process parameters for this comparative analysis are set with a friction factor of 0.1, a feed speed of 0.3 mm/s, and a die rotational speed of 6.28 rad/s.
Influence on Forming Load
The evolution of the forming load during the rolling process for the three dies exhibits a similar single-peak trend, which is characteristic of the through-feed process for cylindrical gears. The load increases during initial penetration, reaches a maximum when the cutting section is fully engaged, stabilizes during finishing, and decreases as the exiting section disengages.
| Beveloid Rolling Die Configuration | Maximum Forming Load (N) | Observation |
|---|---|---|
| Die I ($X_{cut}$: -2→0) | 40,500 | Lowest peak load. |
| Die II ($X_{cut}$: -1→1) | 40,700 | Moderate peak load. |
| Die III ($X_{cut}$: 0→2) | 41,300 | Highest peak load. Lower radial force component during stable forming due to larger positive modification. |
The results clearly indicate that Die I, which starts with a negative modification coefficient, requires the lowest maximum force to form the cylindrical gear. This is advantageous for reducing die stress and equipment capacity requirements.
Influence on Material Flow Velocity
The material flow velocity at different depths from the surface of the cylindrical gear workpiece was tracked. A common pattern was observed: velocity fluctuations were most severe at the surface and dampened significantly at a depth of 2 mm due to constraints from surrounding material. The flow exhibited periodic variations corresponding to the die tooth engagement.
| Evaluation Zone | Die I Max Velocity (mm/s) | Die II Max Velocity (mm/s) | Die III Max Velocity (mm/s) | Note |
|---|---|---|---|---|
| Workpiece Surface (P1-P10) | 25.0 | 25.0 | 21.5 | Dies I & II promote higher surface flow. |
| Depth = 1 mm (P11-P20) | 5.8 | 8.5 | 8.1 | Die I induces significantly lower sub-surface flow. |
| Depth = 2 mm (P21-P30) | 5.4 | 8.5 | 2.8 | Die I shows moderate, stable deep material flow. |
Die I demonstrates a favorable flow pattern: it enables sufficient surface flow for tooth filling while generating lower and potentially more stable flow in the subsurface regions. This can help reduce internal defects and “earing” tendencies in the cylindrical gear.
Influence on Tooth Profile Error
The accuracy of the formed cylindrical gear tooth profile was assessed by comparing it to the theoretical involute profile. The profile error was measured at three cross-sections along the gear axis.
| Cross-Section | Die I Max Error (mm) | Die II Max Error (mm) | Die III Max Error (mm) | Primary Error Location & Comment |
|---|---|---|---|---|
| Section E1 | 0.18 | 0.19 | Largest | Transition fillet for I & II. Die III shows large errors on flank and insufficient tooth height. |
| Section E2 (Center) | 0.15 | 0.22 | 0.23 | Transition fillet for all. Die III errors on flank are much larger. |
| Section E3 | 0.13 | 0.18 | 0.18 | Transition fillet for all. Die III shows large errors on flank. |
Die I consistently produces the cylindrical gear with the smallest tooth profile error across all sections. Die III, with its positive modification progression, results in significant inaccuracies, particularly on the tooth flank and with insufficient addendum height, rendering it unsuitable for precision cylindrical gear forming.
Conclusion of Geometric Analysis: Based on the comprehensive evaluation of forming load, material flow, and profile accuracy, the beveloid rolling die configuration with a modification coefficient progressing from -2 to 0 in the cutting section, 0 in the finishing section, and 0 to -0.5 in the exiting section (Die I) is identified as the optimal design for the through-feed rolling of cylindrical gears.
Process Parameter Optimization for the Optimal Beveloid Die
Having established the optimal geometry for the beveloid rolling die (Die I), the next step is to determine the best operating parameters to minimize forming load and tooth error while maximizing the effective tooth height of the cylindrical gear. An Orthogonal Experimental Design (L9 array) is employed to systematically investigate the influence of three key process parameters: Friction Factor (A), Feed Speed (B), and Die Rotation Speed (C), each at three levels.
| Factor | Level 1 | Level 2 | Level 3 |
|---|---|---|---|
| A: Friction Factor | 0.15 | 0.30 | 0.45 |
| B: Feed Speed (mm/s) | 0.50 | 0.75 | 1.00 |
| C: Die Rotation Speed (rad/s) | 1.34 | 3.14 | 4.94 |
Analysis of Experimental Results
The results for the three evaluation metrics—Effective Tooth Height, Maximum Forming Load, and Maximum Profile Error—from the nine simulation trials are analyzed using range analysis to determine the primary/secondary influence of factors and the optimal level combination.
1. Effective Tooth Height: The target is to achieve the full theoretical addendum height (42 mm in this case). Range analysis indicates that Die Rotation Speed (C) has the greatest influence ($R_C$ = 0.0899), followed by Friction Factor (A) and then Feed Speed (B). The optimal level combination for maximizing tooth height is A1B1C3 (Friction=0.15, Feed=0.5 mm/s, Rotation=4.94 rad/s). This combination ensures sufficient material flow and folding to fill the tooth tip of the cylindrical gear completely.
2. Maximum Forming Load: Minimizing the peak load is crucial for die life and energy consumption. The range analysis shows that Die Rotation Speed (C) is again the most significant factor ($R_C$ = 1608.4 N), with Friction Factor and Feed Speed being less influential. The optimal levels for minimizing load are A1B1C2 (Friction=0.15, Feed=0.5 mm/s, Rotation=3.14 rad/s). A moderate rotation speed balances deformation rate and material flow resistance.
3. Maximum Tooth Profile Error: Achieving dimensional accuracy is paramount. The analysis reveals that Die Rotation Speed (C) is the dominant factor affecting profile error ($R_C$ = 0.1424 mm), followed by Friction Factor. The optimal combination for the most accurate cylindrical gear profile is A1B1C3 (Friction=0.15, Feed=0.5 mm/s, Rotation=4.94 rad/s).
| Optimization Objective | Primary Influence Order (Factor) | Optimal Level Combination (A, B, C) |
|---|---|---|
| Maximize Effective Tooth Height | C > A > B | (1, 1, 3) |
| Minimize Maximum Forming Load | C > A > B | (1, 1, 2) |
| Minimize Maximum Profile Error | C > A > B | (1, 1, 3) |
Determination of Composite Optimal Parameters: A comprehensive analysis weighing all three objectives is required. The combinations A1B1C3 and A1B1C2 are strong candidates. Since Die Rotation Speed (C) is the most influential factor, and the combination A1B1C3 simultaneously optimizes both tooth height and profile accuracy—two critical quality metrics for the cylindrical gear—it is selected as the composite optimum. The slight increase in forming load compared to the C2 level is acceptable given the significant improvements in gear quality. Therefore, the recommended process parameters for use with the optimal beveloid rolling die (Die I) are: Friction Factor = 0.15, Feed Speed = 0.5 mm/s, Die Rotation Speed = 4.94 rad/s.
Experimental Verification
To validate the simulation findings, a physical through-feed rolling experiment was conducted on a custom-built platform. The beveloid rolling die was manufactured according to the Die I specification (cutting: $X$ = -2→0, finishing: $X$ = 0, exiting: $X$ = 0→-0.5). The workpiece was an AL-1100 disk with an initial diameter of Φ80.8 mm and a thickness of 25 mm. The process was run with the optimized parameters: die rotation speed of 4.94 rad/s and an axial feed speed of 0.5 mm/s.
The successfully rolled cylindrical gear was inspected. The formed gear exhibited uniform tooth distribution around the circumference, complete and well-defined tooth profiles, and dimensions meeting the target specifications. The measured average addendum circle radius and dedendum circle radius aligned well with the theoretical values and simulation predictions. This experimental result confirms the feasibility and effectiveness of the proposed geometric design method for the beveloid rolling die and the associated process optimization for the through-feed rolling of cylindrical gears.
Conclusion
This work presents a comprehensive methodology for designing and optimizing the through-feed rolling process for cylindrical gears. The core contribution is the development of a beveloid rolling die with axially varying tooth thickness. The geometric design is derived from the principle of generating a cylindrical gear with a continuously variable profile shift coefficient, leading to precise mathematical models for the cutting, finishing, and exiting sections of the die.
Finite element analysis comparing different modification coefficient ranges demonstrated that a die with a progression from negative to zero modification in the cutting section (e.g., -2 to 0) yields the most favorable outcomes. This configuration for the cylindrical gear rolling process results in the lowest forming load, promotes a controlled material flow pattern that minimizes subsurface deformation, and produces the smallest tooth profile error compared to dies with positive or symmetric modification progressions.
Furthermore, an orthogonal experimental study identified the process parameters that best complement the optimal die geometry. The die rotational speed was found to be the most significant factor influencing forming load, profile error, and effective tooth height of the cylindrical gear. The recommended optimal parameter set is a low friction factor (0.15), a low feed speed (0.5 mm/s), and a relatively high die rotation speed (4.94 rad/s). Experimental trials conducted with the designed beveloid die and these parameters successfully produced cylindrical gears with excellent geometric integrity, thereby validating the entire design and optimization framework. This approach provides a valuable theoretical and practical foundation for improving the efficiency and quality of the cylindrical gear through-feed rolling process.