In my work on automotive drivetrain components, I focused on the rotary swaging of a hollow gear shaft produced from a forged gear blank. The motivation was straightforward: vehicle lightweighting has become one of the most practical routes to reduce fuel consumption and emissions, and a hollow gear shaft can replace a solid shaft while maintaining or even improving torsional performance. The forged gear blank used in my study was made from 27MnCr5 steel, and the final shaping operation was a precision rotary swaging process performed on a four-hammer radial forging machine. I combined orthogonal experiment design with Forge simulation to identify a robust parameter window for the forged gear blank, and I then validated the optimized process through microstructural observation and microhardness testing.
The concept of a forged gear blank is central to this investigation. A forged gear blank is not merely a preform; it defines the initial grain flow, the distribution of alloying elements, and the thermal history that later rotary swaging must exploit. When I refer to the forged gear blank, I mean the starting workpiece that already possesses a favorable wrought structure and near-net geometry. The rotary swaging operation then refines that structure, reduces the wall thickness, and imparts a fibrous architecture along the axial direction. In my experiments, the forged gear blank was heated to a controlled initial temperature and fed axially between rotating hammers and a mandrel. The hammers applied high-frequency radial blows, while the workpiece rotated and advanced. This combination produced a complex strain path that I needed to understand and optimize.
My study addressed two dominant defects in the rotary swaging of the forged gear blank: material folding and spiral ridge marks. Material folding appears when the radial reduction per pass is too aggressive relative to the axial feed, causing surface material to fold back into the workpiece. Spiral ridge marks, often called spiral spinal marks, arise from an unfavorable interaction between hammer velocity, rotational speed, and axial feed. Both defects reduce the effective load-bearing area, create stress concentrations, and can lead to premature fatigue failure. Therefore, I treated folding count and spiral ridge depth as the primary quality responses in my orthogonal experiment.
Material and forged gear blank characteristics
The material chosen for the forged gear blank was 27MnCr5 steel. This alloy is widely used in transmission and drivetrain applications because it combines moderate hardenability, good toughness, and favorable forging behavior. I verified the chemical composition of the as-received forged gear blank before any rotary swaging trials. The composition is summarized in Table 1. The carbon content provides sufficient hardness after heat treatment, while manganese, chromium, and molybdenum contribute to hardenability and wear resistance. The low phosphorus and sulfur levels help avoid hot cracking and inclusion-related damage during radial forging.
| Element | C | Si | Mn | P | S | Cr | Mo | Fe |
|---|---|---|---|---|---|---|---|---|
| Range | 0.38–0.45 | 0.17–0.37 | 0.50–0.80 | ≤0.035 | ≤0.035 | 0.90–1.20 | 0.15–0.25 | Balance |
The forged gear blank geometry included a central bore and stepped outer diameters. The stepped geometry is important because the radial reduction and wall thickness change along the axis. During rotary swaging, the mandrel supports the inner surface while the hammers compress the outer surface. The material flow is therefore constrained both radially and axially. I observed that the second step of the forged gear blank experienced a larger total deformation than the first step, which later proved essential in explaining the microhardness distribution.
Rotary swaging principle and process variables
I used an imported GFM SKK-10 precision radial forging machine for the rotary swaging trials. The machine operated with multiple hammers arranged around the workpiece. The forged gear blank rotated continuously while being fed axially through the hammer zone. A mandrel was inserted into the central bore to control the inner diameter and to prevent collapse of the hollow section. The hammer radial reduction, hammer radial velocity, axial relative feed, rotational speed, and initial temperature were the five factors I selected for optimization.
The axial relative feed per revolution can be expressed as
$$ f_a = \frac{v_a}{n_b} $$
where \(f_a\) is the relative axial feed per revolution, \(v_a\) is the axial feed velocity, and \(n_b\) is the rotational speed of the forged gear blank. A small relative feed allows more hammer blows per unit axial length and generally produces a smoother surface, but it reduces productivity. A large relative feed can cause spiral ridge marks and incomplete deformation. The radial reduction ratio is another key variable:
$$ r = \frac{D_0 – D_f}{D_0} \times 100\% $$
where \(D_0\) is the initial outer diameter of the forged gear blank and \(D_f\) is the final outer diameter after swaging. The effective strain imposed on the hollow section can be approximated by
$$ \varepsilon_{\text{eff}} = \ln\left(\frac{A_0}{A_f}\right) $$
where \(A_0\) and \(A_f\) are the initial and final cross-sectional areas of the wall. In stepped regions, the effective strain varies along the axis, which leads to non-uniform grain refinement and hardness. I used these relationships to interpret both the simulation results and the physical measurements.
Orthogonal experiment design
Because a full factorial experiment over five factors at four levels would require 1024 runs, I adopted an orthogonal array. This allowed me to evaluate the main effects with only 16 trials. The factors and levels are shown in Table 2. Factor A was the initial temperature of the forged gear blank, factor B was the axial relative feed, factor C was the rotational speed of the forged gear blank, factor D was the hammer radial reduction, and factor E was the hammer radial velocity.
| Level | A: Initial temperature (°C) | B: Axial relative feed (mm) | C: Rotational speed (r·min⁻¹) | D: Hammer radial reduction (mm) | E: Hammer radial velocity (mm·s⁻¹) |
|---|---|---|---|---|---|
| 1 | 750 | 0.05 | 50 | 2.0 | 7.5 |
| 2 | 780 | 0.10 | 60 | 2.5 | 10.0 |
| 3 | 800 | 0.30 | 70 | 3.0 | 12.0 |
| 4 | 830 | 0.50 | 76 | 3.5 | 15.0 |
I used Forge software to simulate each run. The forged gear blank was modeled as a deformable body, the mandrel and hammers were modeled as rigid bodies, and the thermal boundary conditions included contact heat transfer, convection, and radiation. The simulation output was inspected for folding events and spiral ridge patterns. To make the results quantitative, I assigned discrete scores to each defect. For folding, a score of 0 meant no significant folding or only minor mesh-overlap artifacts, a score of 1 meant shallow folding, and a score of 2 meant deep folding. For spiral ridge depth, a score of 0 meant no visible ridge, a score of 1 meant a shallow ridge, and a score of 2 meant a pronounced ridge. The lower the score, the better the process condition for the forged gear blank.
| Run | A | B | C | D | E | Process condition | Folding score \(n\) | Spiral ridge score \(h\) |
|---|---|---|---|---|---|---|---|---|
| 1 | 1 | 1 | 1 | 1 | 1 | A1B1C1D1E1 | 0 | 0 |
| 2 | 1 | 2 | 2 | 2 | 2 | A1B2C2D2E2 | 1 | 1 |
| 3 | 1 | 3 | 3 | 3 | 3 | A1B3C3D3E3 | 2 | 1 |
| 4 | 1 | 4 | 4 | 4 | 4 | A1B4C4D4E4 | 2 | 2 |
| 5 | 2 | 1 | 2 | 3 | 4 | A2B1C2D3E4 | 2 | 0 |
| 6 | 2 | 2 | 1 | 4 | 3 | A2B2C1D4E3 | 2 | 1 |
| 7 | 2 | 3 | 4 | 1 | 2 | A2B3C4D1E2 | 0 | 1 |
| 8 | 2 | 4 | 3 | 2 | 1 | A2B4C3D2E1 | 1 | 2 |
| 9 | 3 | 1 | 3 | 4 | 2 | A3B1C3D4E2 | 2 | 2 |
| 10 | 3 | 2 | 4 | 3 | 1 | A3B2C4D3E1 | 1 | 1 |
| 11 | 3 | 3 | 1 | 2 | 4 | A3B3C1D2E4 | 2 | 1 |
| 12 | 3 | 4 | 2 | 1 | 3 | A3B4C2D1E3 | 2 | 2 |
| 13 | 4 | 1 | 4 | 2 | 3 | A4B1C4D2E3 | 0 | 0 |
| 14 | 4 | 2 | 3 | 1 | 4 | A4B2C3D1E4 | 0 | 0 |
| 15 | 4 | 3 | 2 | 4 | 1 | A4B3C2D4E1 | 2 | 1 |
| 16 | 4 | 4 | 1 | 3 | 2 | A4B4C1D3E2 | 2 | 2 |
Range analysis and factor significance
For each factor and level, I calculated the mean response. If \(k_{ij}\) is the mean response for factor \(i\) at level \(j\), then
$$ k_{ij} = \frac{1}{s} \sum_{r=1}^{s} y_{ijr} $$
where \(s\) is the number of runs at that level and \(y_{ijr}\) is the observed score. The range of each factor is
$$ R_i = \max_j(k_{ij}) – \min_j(k_{ij}) $$
A larger range indicates a stronger influence on the quality response. Table 4 and Table 5 summarize the range analysis for folding score and spiral ridge score. For folding, the factor order was \(D > C > A = B > E\). For spiral ridge depth, the factor order was \(B > A = D = E > C\). These orders guided my selection of the final process window.
| Level | A | B | C | D | E |
|---|---|---|---|---|---|
| \(k_1\) | 1.25 | 1.00 | 1.50 | 0.50 | 1.00 |
| \(k_2\) | 1.25 | 1.00 | 1.75 | 1.00 | 1.25 |
| \(k_3\) | 1.75 | 1.50 | 1.25 | 1.75 | 1.50 |
| \(k_4\) | 1.00 | 1.75 | 0.75 | 2.00 | 1.50 |
| Range \(R\) | 0.75 | 0.75 | 1.00 | 1.50 | 0.50 |
| Rank | 3 | 3 | 2 | 1 | 4 |
| Level | A | B | C | D | E |
|---|---|---|---|---|---|
| \(k_1\) | 1.00 | 0.50 | 1.00 | 0.75 | 1.00 |
| \(k_2\) | 1.00 | 0.75 | 1.00 | 1.00 | 1.50 |
| \(k_3\) | 1.50 | 1.00 | 1.25 | 1.00 | 1.00 |
| \(k_4\) | 0.75 | 2.00 | 1.00 | 1.50 | 0.75 |
| Range \(R\) | 0.75 | 1.50 | 0.25 | 0.75 | 0.75 |
| Rank | 2 | 1 | 4 | 2 | 2 |
For a defect score, smaller is better. Therefore, I selected the level with the smallest mean for each factor while also considering practical constraints. For factor A, the fourth level gave the lowest mean for both responses, so I chose an initial temperature of 830 °C. For factor B, levels 1 and 2 were similar for folding, but level 1 was better for spiral ridge depth, so I chose 0.05 mm. For factor C, the fourth level was best for folding and tied for best for spiral ridge depth, so I chose 76 r·min⁻¹. For factor D, the first level was best for both responses, so I chose a radial reduction of 2 mm. For factor E, the first level was best for folding and tied for best for spiral ridge depth, so I chose 7.5 mm·s⁻¹. The resulting optimized condition was A4B1C4D1E1. This condition was used for the physical validation of the forged gear blank.
| Parameter | Symbol | Optimized value |
|---|---|---|
| Initial temperature | A | 830 °C |
| Axial relative feed | B | 0.05 mm |
| Rotational speed | C | 76 r·min⁻¹ |
| Hammer radial reduction | D | 2.0 mm |
| Hammer radial velocity | E | 7.5 mm·s⁻¹ |
I also constructed a composite score to check the trade-off between folding and spiral ridges. A simple weighted objective is
$$ S = w_n n + w_h h $$
where \(w_n\) and \(w_h\) are weights. For a balanced quality target, I set \(w_n = w_h = 0.5\). Under that composite score, the optimized condition A4B1C4D1E1 still ranked among the best combinations, confirming that the parameter set was not overly biased toward one defect type.
Forge simulation insights for the forged gear blank
The Forge simulation helped me visualize the material flow within the forged gear blank. In the early stage of each hammer blow, the outer surface experienced a localized radial compression. As the workpiece rotated, the deformation zone moved around the circumference. With a small axial feed, the deformation zones overlapped more completely, which reduced the formation of spiral ridges. With a large axial feed, the deformation zones were separated, leaving an unrefined helical band. The simulation also showed that a radial reduction of 2 mm was sufficient to close the internal porosity of the forged gear blank without causing surface folding. When the radial reduction exceeded 3 mm, the surface layer tended to fold over the mandrel step, especially near the second step.
The thermal field was equally important. The initial temperature of the forged gear blank affected the flow stress and the dynamic recrystallization behavior. A higher initial temperature reduced the flow stress, which lowered the hammer load and improved material flow. However, an excessively high temperature can cause grain growth. In my simulations, 830 °C provided a favorable balance: the flow stress was low enough to avoid folding, but the temperature was not so high that the final grain size became coarse. The constitutive response of the forged gear blank can be represented in a general form as
$$ \sigma = K \varepsilon^n \dot{\varepsilon}^m \exp\left(-\frac{Q}{RT}\right) $$
where \(\sigma\) is the flow stress, \(\varepsilon\) is the true strain, \(\dot{\varepsilon}\) is the strain rate, \(T\) is the absolute temperature, \(Q\) is the activation energy, \(R\) is the gas constant, and \(K\), \(n\), and \(m\) are material constants. This equation explains why increasing temperature lowers the flow stress and why increasing strain rate raises it. In rotary swaging, the strain rate is high because the hammer contact time is short. Therefore, the forged gear blank must be hot enough to avoid cracking, but the hammer velocity must not be so high that adiabatic heating or surface damage occurs.
The effective strain distribution from the simulation followed an exponential decay from the outer surface toward the inner surface. A useful approximation is
$$ \varepsilon_{\text{eff}}(r) = \varepsilon_{\text{surf}} \exp\left[-\alpha \left(\frac{R_0 – r}{t_w}\right)\right] $$
where \(\varepsilon_{\text{surf}}\) is the surface effective strain, \(R_0\) is the outer radius, \(r\) is the radial coordinate, \(t_w\) is the wall thickness, and \(\alpha\) is a decay coefficient. This distribution is consistent with the microhardness results: the surface and near-surface regions experienced more deformation, while the mid-wall region experienced less. The second step had a larger total reduction, so its surface strain was higher than that of the first step.
Microstructural validation of the forged gear blank
After rotary swaging with the optimized parameters, I sectioned the forged gear blank and examined four distinct regions on the cross-section. I used scanning electron microscopy to observe the grain structure. The general trend was clear: as the deformation increased, the grains became more refined. From the surface toward the interior, the grain refinement first increased and then gradually weakened. This means that the most refined grains were not always exactly at the outermost surface; in some regions, the outermost layer retained a small fraction of coarse grains because of friction, chilling, or incomplete recrystallization.
In the first region, I observed both deformed grains and remnant original grains. The grain size decreased from the surface toward the interior. At higher magnification, the deformed grains were elongated along the axial direction. In the second region, the stepped cross-section showed a similar gradient, but the elongated grains were more pronounced. The material had developed a fibrous architecture, which is desirable for torsional and bending loads. In the third region, I found a fibrous band about 50 μm below the surface. The fiber orientation was parallel to the axial drawing direction of the forged gear blank. In the fourth region, where the material flowed around the mandrel step, the grains were highly refined and the structure was very dense. The mandrel step obstructed the axial flow, forcing the material to turn and consolidate.
The formation of fibrous grains can be described by the geometric thinning of grains during plastic deformation:
$$ \lambda = \frac{L_f}{L_0} = \exp(\varepsilon_{\text{eff}}) $$
where \(\lambda\) is the elongation ratio, \(L_0\) is the initial grain length, and \(L_f\) is the final grain length. For a forged gear blank with a well-controlled initial microstructure, this elongation produces a continuous fiber flow that improves fatigue resistance. The grain refinement also follows the Hall–Petch relationship:
$$ \sigma_y = \sigma_0 + k_y d^{-1/2} $$
where \(\sigma_y\) is the yield strength, \(\sigma_0\) is the friction stress, \(k_y\) is the strengthening coefficient, and \(d\) is the average grain diameter. A similar relation can be written for hardness:
$$ HV = H_0 + k_H d^{-1/2} $$
Thus, the finer grains in the more heavily deformed regions should produce higher hardness. This expectation matched my microhardness measurements.
| Region | Location | Dominant feature | Grain refinement trend |
|---|---|---|---|
| 1 | First step surface | Deformed grains plus remnant original grains | Refinement increases from surface to interior |
| 2 | Second step surface | Elongated grains and fibrous flow | Stronger refinement than region 1 |
| 3 | Near-surface band | Fiber orientation parallel to axial direction | Fine fibrous structure |
| 4 | Mandrel step flow zone | Dense, highly refined fiber structure | Maximum refinement and densification |
Microhardness response of the forged gear blank
I measured microhardness on the RD–TD plane of the forged gear blank. The load was 100 g and the dwell time was 10 s. At each location, I took six equally spaced points along the length direction. To improve reliability, I also took two additional points on each side of the width direction and averaged the three measurements. The average hardness at a location is
$$ HV_{\text{avg}} = \frac{1}{3}\sum_{i=1}^{3} HV_i $$
The hardness curves showed a clear gradient from the inner surface to the outer surface. The inner surface, which was in contact with the mandrel, often had higher hardness than the mid-wall region. Two reasons explain this. First, the material flow near the mandrel was intense because the mandrel constrained the inner surface and forced the metal to flow along the axial direction. Second, the mandrel acted as a heat sink, and the cooling effect suppressed grain growth. Therefore, the inner surface retained a finer structure and higher hardness.
The first step and the shaping segment showed different hardness levels. The shaping segment had a higher hardness than the stepped region because the shaping pass imposed additional deformation and densification. Similarly, the second step and its shaping segment showed the same trend: the shaping segment had higher hardness than the stepped region. More importantly, the second step had a higher surface hardness than the first step. This is consistent with the larger total deformation at the second step. The difference can be expressed as
$$ \Delta HV = HV_{\text{step2}} – HV_{\text{step1}} > 0 $$
The hardness variation also reflects the non-uniform strain distribution. If the deformation were perfectly uniform, the hardness would be constant along the radial direction. Instead, I observed a gradient, which means that the forged gear blank underwent heterogeneous deformation. This heterogeneity is not necessarily harmful; in fact, a harder surface and a tougher interior can be beneficial for gear shafts. However, excessive heterogeneity can lead to residual stresses and dimensional instability. Therefore, the optimized process aimed to control, not eliminate, the strain gradient.
| Region | Relative hardness | Main reason |
|---|---|---|
| First step, inner surface | High | Mandrel contact, intense flow, cooling |
| First step, mid-wall | Lower | Less effective strain |
| First step, shaping segment | Higher than step | Additional densification |
| Second step, inner surface | Highest | Larger total reduction |
| Second step, mid-wall | Intermediate | Moderate strain |
| Second step, shaping segment | Higher than step | Final shaping and densification |
Discussion of parameter effects
The orthogonal analysis and physical validation allowed me to rank the importance of each parameter for the forged gear blank. The hammer radial reduction had the strongest effect on folding. When the radial reduction was too large, the surface layer could not flow fast enough, and it folded over itself. The rotational speed was the second most important factor for folding because it controlled the number of hammer blows per unit time. A higher rotational speed, combined with a small axial feed, produced a more uniform deformation field. The initial temperature and axial feed had similar effects on folding. A higher temperature improved ductility, while a smaller axial feed allowed more overlaps. The hammer radial velocity had the smallest effect on folding, but it still influenced the surface finish.
For spiral ridge depth, the axial feed was the dominant factor. A small axial feed reduced the helical pitch and allowed the deformation zones to overlap, which smoothed the surface. The initial temperature, radial reduction, and hammer radial velocity had moderate effects. The rotational speed had the weakest effect on spiral ridges, although it interacted with the axial feed. These interactions are important because a parameter that is beneficial alone may be harmful when combined with another parameter. For example, a high rotational speed with a large axial feed can still produce spiral marks because the axial advance per revolution is too large.
| Parameter | Effect on folding | Effect on spiral ridges | Recommended value |
|---|---|---|---|
| Initial temperature | Moderate | Moderate | 830 °C |
| Axial relative feed | Moderate | Strong | 0.05 mm |
| Rotational speed | Strong | Weak | 76 r·min⁻¹ |
| Hammer radial reduction | Strong | Moderate | 2.0 mm |
| Hammer radial velocity | Weak | Moderate | 7.5 mm·s⁻¹ |
I also considered the interaction between the forged gear blank geometry and the mandrel. The mandrel diameter determines the inner wall support. If the mandrel is too small, the hollow section may collapse or wrinkle. If the mandrel is too large, the wall thickness reduction is limited. In my trials, the mandrel was matched to the initial bore of the forged gear blank so that the inner surface remained in contact during the entire swaging pass. This contact is essential for producing a dense inner layer and for preventing internal defects. The simulation showed that the contact pressure was highest near the step transitions, which explains why the fibrous structure was most pronounced there.
The temperature history also affected the final properties. The forged gear blank started at 830 °C, but the surface temperature dropped because of contact with the cooler hammers and mandrel. The interior remained hotter. This thermal gradient influenced dynamic recovery and recrystallization. In regions where the temperature remained high and the strain was large, recrystallization produced fine equiaxed grains. In regions where the temperature dropped quickly, the deformed structure was retained as elongated grains. The mixed structure I observed is therefore a natural consequence of the coupled thermal and mechanical fields.
From a production standpoint, the optimized parameters also improved stability. A small axial feed of 0.05 mm and a rotational speed of 76 r·min⁻¹ produced a smooth surface and reduced the risk of folding. A radial reduction of 2 mm per pass was sufficient to refine the forged gear blank without overloading the hammers. The hammer radial velocity of 7.5 mm·s⁻¹ provided a controlled impact that avoided excessive adiabatic heating. These settings are practical for a GFM SKK-10 machine because they fall within the machine capacity and do not require unusual tooling.
Process optimization logic and formula summary
The optimization logic can be summarized mathematically. Let the quality response vector be
$$ \mathbf{y} = \begin{bmatrix} n \\ h \end{bmatrix} $$
where \(n\) is the folding score and \(h\) is the spiral ridge score. The goal is to minimize both responses. The factor-level mean matrix is
$$ \mathbf{K} = \begin{bmatrix} k_{A1} & k_{B1} & k_{C1} & k_{D1} & k_{E1} \\ k_{A2} & k_{B2} & k_{C2} & k_{D2} & k_{E2} \\ k_{A3} & k_{B3} & k_{C3} & k_{D3} & k_{E3} \\ k_{A4} & k_{B4} & k_{C4} & k_{D4} & k_{E4} \end{bmatrix} $$
For each column, the best level is the one with the minimum mean. The range vector is
$$ \mathbf{R} = \begin{bmatrix} R_A & R_B & R_C & R_D & R_E \end{bmatrix} $$
The factor ranking follows the descending order of \(\mathbf{R}\). For folding, the ranking was \(D > C > A = B > E\). For spiral ridges, the ranking was \(B > A = D = E > C\). The final decision balanced both rankings and practical machine limits. The optimized parameter vector is
$$ \mathbf{x}^* = \begin{bmatrix} 830 & 0.05 & 76 & 2.0 & 7.5 \end{bmatrix} $$
This vector represents the best combination I found for the forged gear blank. It is not simply the best level for each factor in isolation; it is the result of a compromise between defect suppression, grain refinement, and process feasibility.
Quality implications for the forged gear blank
The rotary swaging process improved the forged gear blank in several ways. First, it reduced the wall thickness and created a hollow section, which lowered the mass. Second, it refined the grains, especially near the inner and outer surfaces. Third, it produced a fibrous flow line that follows the shaft contour. Fourth, it increased the density of the material by closing small internal voids. Fifth, it created a favorable hardness gradient, with harder surfaces and a tougher core. These benefits are directly related to the needs of a gear shaft, which must transmit torque, resist bending, and endure cyclic loading.
The folding and spiral ridge defects were minimized by the optimized parameters. In the trial forged gear blank, I did not observe deep folding or pronounced spiral ridges. The surface was smooth enough for subsequent machining or finishing. The microstructure was dense and fibrous. The microhardness followed the expected trend. Therefore, I concluded that the optimized rotary swaging process is suitable for producing a high-quality hollow gear shaft from a forged gear blank.
Practical recommendations for industrial implementation
For industrial implementation, I recommend the following. First, control the initial temperature of the forged gear blank within a narrow band around 830 °C. Temperature variation of more than 20 °C can change the flow stress and affect defect formation. Second, keep the axial relative feed at 0.05 mm. This small feed requires more passes or slower axial advance, but it produces a better surface and reduces spiral ridges. Third, use a rotational speed of 76 r·min⁻¹. This speed provides enough hammer blows per revolution to overlap the deformation zones. Fourth, limit the hammer radial reduction to 2 mm per pass. A larger reduction increases the risk of folding. Fifth, use a hammer radial velocity of 7.5 mm·s⁻¹. This velocity is high enough to deform the material but low enough to avoid excessive impact heating. Sixth, maintain good alignment between the mandrel and the forged gear blank. Misalignment causes wall thickness variation and can lead to internal defects. Seventh, monitor the mandrel contact pressure. A sudden drop in pressure may indicate wall collapse or folding. Eighth, verify the grain flow after a trial run. The fiber orientation should be parallel to the axial direction, and the grains should be finer at the surface than in the core.
| Parameter | Optimized value | Acceptable range | Primary benefit |
|---|---|---|---|
| Initial temperature | 830 °C | 810–850 °C | Lower flow stress, controlled grain growth |
| Axial relative feed | 0.05 mm | 0.03–0.08 mm | Reduced spiral ridges |
| Rotational speed | 76 r·min⁻¹ | 70–80 r·min⁻¹ | Uniform deformation |
| Hammer radial reduction | 2.0 mm | 1.5–2.5 mm | Defect-free reduction |
| Hammer radial velocity | 7.5 mm·s⁻¹ | 7.0–10.0 mm·s⁻¹ | Controlled impact and surface finish |
Conclusion
I optimized the rotary swaging process for a hollow gear shaft made from a forged gear blank of 27MnCr5 steel. Using an orthogonal experiment with five factors and four levels, I evaluated the influence of initial temperature, axial relative feed, rotational speed, hammer radial reduction, and hammer radial velocity on folding and spiral ridge defects. The range analysis showed that hammer radial reduction and rotational speed were the most important factors for folding, while axial relative feed was the most important factor for spiral ridges. The optimized parameter combination was an initial temperature of 830 °C, an axial relative feed of 0.05 mm, a rotational speed of 76 r·min⁻¹, a hammer radial reduction of 2.0 mm, and a hammer radial velocity of 7.5 mm·s⁻¹.
The physical validation confirmed that the optimized process produced a forged gear blank with refined grains, fibrous flow lines, and improved density. Grain refinement was strongest in the regions with the highest effective strain, and it decreased from the surface toward the interior. The second step experienced a larger deformation than the first step, which led to a higher surface microhardness. The inner surface near the mandrel also showed higher hardness because of intense material flow and local cooling. These results demonstrate that rotary swaging is an effective near-net-shape process for producing hollow gear shafts from a forged gear blank. The combination of orthogonal design, Forge simulation, and physical testing provided a reliable path to optimize the process and to understand the relationship between process parameters, microstructure, and hardness.
In future work, I would extend the study to include fatigue testing and residual stress measurement. I would also examine the effect of different mandrel materials and coatings on the inner surface quality of the forged gear blank. In addition, I would explore closed-loop control of the axial feed and rotational speed to maintain the optimized conditions throughout the entire swaging pass. These steps would further strengthen the industrial applicability of the process and help ensure consistent quality in mass production of hollow gear shafts from forged gear blanks.