In the realm of modern manufacturing, the advancement of automation within the casting sector is pivotal. Precision investment casting stands as a prominent near-net-shape advanced process, renowned for its ability to produce components with intricate geometries, excellent surface finish, and minimal need for subsequent machining. This makes precision investment casting indispensable across various high-tech industries, including aerospace, automotive, and medical devices. However, the traditional manual shell-making process, which involves repeated dipping, sanding, and drying of wax patterns, presents significant challenges. These include low production efficiency, harsh working environments characterized by dust and fumes, and high labor intensity for workers. To address these issues and align with the trends of industrial automation and smart manufacturing, I embarked on designing a specialized robotic manipulator to automate the shell-making process in precision investment casting. This manipulator aims to replace human labor in handling wax pattern assemblies, thereby enhancing productivity, improving workplace safety, and reducing operational costs. The core of this work focuses on the structural design and optimization of the manipulator’s end-effector—the gripper mechanism—to ensure reliable performance, sufficient strength, and lightweight construction.
The shell-making process in precision investment casting begins with the creation of wax patterns, which are then assembled onto a central wax sprue or a support plate, often referred to as a top cover plate. This assembly, known as a wax tree, is subsequently subjected to a series of coating operations to build up a ceramic shell. The process typically involves repetitive cycles of dipping the assembly into a ceramic slurry, sprinkling it with refractory sand, and allowing it to dry. Automating this sequence requires a manipulator capable of securely grasping the top cover plate of the wax tree, which usually has standardized interface features, and maneuvering it through the various stations. The primary design requirements for the manipulator’s gripper were derived from these operational needs. It must be able to grasp a cylindrical feature on the top cover plate with a diameter of Φ=20 mm and a height of H=20 mm. Furthermore, it must handle a total load, including the wax tree assembly, not exceeding 10 kg. Key design considerations included ensuring a large enough gripping force for reliable handling, ease of positioning and clamping, sufficient structural strength to withstand operational loads, and, critically, achieving a lightweight design to reduce inertia and improve dynamic performance without compromising integrity.
The designed end-effector assembly for the shell-making manipulator consists of several key components: a small arm, a push rod, the gripper fingers, a fixed slide rail, connecting links, cylindrical pins, and a linear actuator. The linear motion is provided by a pneumatic cylinder, which actuates the push rod. The extension and retraction of this push rod drive the connecting links, causing the small arm to slide along the fixed rail. The gripper fingers are welded to the small arm, so the arm’s movement directly translates into the opening and closing of the fingers. When the cylinder extends, the fingers open; when it retracts, the fingers close. The contact surfaces of the gripper fingers are designed as V-blocks. This V-shaped profile offers excellent self-centering capability for the cylindrical feature on the cover plate, ensuring stable and repeatable grasping even with minor part or assembly tolerances. The mechanical advantage or force amplification of this linkage mechanism is crucial for efficiency. A schematic of the mechanism can be represented with four revolute joints O1, O2, O3, O4. The gripping force \(F_{grip}\) and the driving force \(F_{drive}\) from the cylinder are related through the geometry. Let the angle between links be \(\alpha\). From static equilibrium analysis, the relationship is derived as:
$$F_{drive} = 2 F_{grip} \sin \alpha \cos \alpha$$
The force amplification ratio \(\tau\) is therefore:
$$\tau = \frac{F_{grip}}{F_{drive}} = \frac{1}{2 \sin \alpha \cos \alpha}$$
For a typical design angle, this provides a favorable mechanical advantage, meaning a relatively small cylinder force can generate a substantial gripping force. The required gripping force must overcome the weight of the workpiece and dynamic inertial forces during acceleration and deceleration. It is calculated with a safety factor:
$$F_{grip} \geq K_1 K_2 K_3 m g$$
Where \(K_1=1.2\) is the safety coefficient, \(K_2=1.5\) is the working condition coefficient, \(K_3=4\) is the orientation coefficient (accounting for friction between the gripper and the plate), \(m=10\) kg is the mass, and \(g=9.8 \, \text{m/s}^2\) is gravity. This yields \(F_{grip} \geq 705.6 \, \text{N}\). A design value of \(F_{grip} = 750 \, \text{N}\) was selected. Using the force relation with a specific angle \(\alpha\), the required cylinder drive force was determined to be approximately 634 N, resulting in a force amplification ratio \(\tau \approx 1.18\), indicating an efficient mechanism design.
To validate the structural integrity of the initial gripper design, a detailed finite element analysis (FEA) was conducted. A three-dimensional solid model of the gripper was created using SolidWorks and then imported into ANSYS Workbench for static structural analysis. The gripper material was specified as stainless steel, which offers a good balance of strength, corrosion resistance, and manufacturability—essential for the often humid and chemically active environment of precision investment casting foundries. The material properties assigned were: density \(\rho = 7850 \, \text{kg/m}^3\), Young’s modulus \(E = 200 \, \text{GPa}\), Poisson’s ratio \(\nu = 0.3\), and yield strength \(\sigma_y = 450 \, \text{MPa}\). The mesh was generated with a refined element size of 0.5 mm for critical areas, resulting in a high-quality mesh with over 150,000 elements and 640,000 nodes. The boundary conditions simulated the actual loading scenario: a fixed support was applied at the interface where the gripper is welded to the small arm, and a force of 750 N was applied on the inner V-surface of the gripper fingers, representing the reaction force from the gripped cover plate. The results of the static analysis are summarized below:
| Parameter | Value |
|---|---|
| Maximum Total Deformation | 0.0055 mm |
| Maximum Equivalent (von-Mises) Stress | 53.357 MPa |
| Mass of Initial Gripper Design | 0.0928 kg |
The maximum deformation was negligible, occurring at the base of the gripper, and would have no practical impact on positioning accuracy. The maximum stress was well below the material’s yield strength, giving a safety factor of \(n = 450 / 53.357 \approx 8.4\), which is conservative. While this confirmed the design was safe, it also indicated significant potential for weight reduction through structural optimization—a key goal for improving the dynamic response and reducing the load on the manipulator’s preceding joints.
To systematically reduce mass while maintaining performance, a topology optimization study was performed. Topology optimization is a computational method that determines the optimal material distribution within a given design space under specified loads and constraints. The goal is to minimize compliance (maximize stiffness) or other objectives subject to a volume constraint. The mathematical formulation for minimizing compliance under a volume constraint, using the variable density method (SIMP), can be expressed as:
$$
\begin{aligned}
& \min: C(s) = U^T K U = \sum_{e=1}^{N} (s_e)^p u_e^T k_0 u_e \\
& \text{subject to: } \frac{V(s)}{V_0} \leq f \\
& \quad \quad \quad \quad F = K U \\
& \quad \quad \quad \quad 0 < s_{min} \leq s_e \leq 1, \quad e = 1,…,N
\end{aligned}
$$
Here, \(C\) is compliance, \(U\) is the displacement vector, \(K\) is the global stiffness matrix, \(s_e\) is the pseudo-density of element \(e\), \(p\) is the penalty factor (typically 3), \(u_e\) and \(k_0\) are element displacement and stiffness matrix, \(V(s)\) is the volume of the structure, \(V_0\) is the volume of the design space, and \(f\) is the prescribed volume fraction. In ANSYS Workbench, the topology optimization module was used with the static structural results as input. The objective was to minimize compliance with a volume reduction target of 40% (i.e., retain 60% of the material). The design space was the entire gripper volume, excluding non-design regions like mounting holes and contact surfaces. The resulting topology optimization density plot showed regions of high material necessity (to be kept) and low material contribution (potential removal). Interpreting this contour plot, the areas suggested for removal were primarily from the inner non-load-bearing webs and some sections of the side walls of the gripper body. This provided a conceptual guide for redesign.
Guided by the topology optimization results, a parametric redesign of the gripper was undertaken. Key dimensions were identified as design variables for further refinement through Response Surface Methodology (RSM). RSM is a collection of statistical techniques for constructing approximate models (meta-models) of system performance based on a set of designed experiments, which can then be used for optimization. Three dimensions were chosen as input variables: the thickness of the gripper’s main body (\(DS\_D1\)), the width of a side rib (\(DS\_D2\)), and the height of a cut-out section (\(DS\_D3\)). Their initial values and ranges for optimization are listed below:
| Design Variable | Initial Value (mm) | Lower Bound (mm) | Upper Bound (mm) |
|---|---|---|---|
| DS_D1 (Body Thickness) | 20.0 | 10.5 | 21.0 |
| DS_D2 (Rib Width) | 10.0 | 6.0 | 11.0 |
| DS_D3 (Cut-out Height) | 22.0 | 18.0 | 23.0 |
The optimization problem was formally defined as a lightweight design with constraints on performance:
$$
\begin{aligned}
& \min: F(x) = m(x) \quad \text{(Mass of the gripper)} \\
& \text{subject to: } \sigma_{max}(x) \leq 180 \, \text{MPa} \\
& \quad \quad \quad \quad \delta_{max}(x) \leq 0.2 \, \text{mm} \\
& \quad \quad \quad \quad x_{lb} \leq x \leq x_{ub}
\end{aligned}
$$
The stress constraint was set more aggressively than the yield strength to ensure a good safety margin after optimization, and the deformation constraint was set to a reasonable limit for positioning accuracy. A Central Composite Design (CCD) was used to sample the design space, generating 15 design points. For each point, a static FEA was automatically performed to compute the mass, maximum stress, and maximum deformation. The results are tabulated below:
| Run | DS_D3 (mm) | DS_D2 (mm) | DS_D1 (mm) | Mass (kg) | Max Deformation (mm) | Max Stress (MPa) |
|---|---|---|---|---|---|---|
| 1 | 20.835 | 6.400 | 15.639 | 0.0625 | 0.0072 | 106.06 |
| 2 | 22.457 | 8.358 | 14.521 | 0.0634 | 0.0077 | 112.58 |
| 3 | 21.209 | 6.984 | 17.939 | 0.0739 | 0.0062 | 93.03 |
| 4 | 21.729 | 7.901 | 16.621 | 0.0713 | 0.0067 | 99.30 |
| 5 | 18.279 | 7.157 | 11.591 | 0.0482 | 0.0099 | 136.05 |
| 6 | 18.715 | 9.546 | 17.087 | 0.0764 | 0.0067 | 95.33 |
| 7 | 20.070 | 8.846 | 20.128 | 0.0889 | 0.0056 | 83.52 |
| 8 | 22.970 | 10.759 | 12.662 | 0.0600 | 0.0092 | 129.44 |
| 9 | 19.099 | 9.438 | 11.422 | 0.0511 | 0.0100 | 137.41 |
| 10 | 19.889 | 10.433 | 19.604 | 0.0905 | 0.0058 | 86.70 |
| 11 | 20.893 | 8.753 | 18.477 | 0.0817 | 0.0061 | 89.67 |
| 12 | 18.629 | 6.649 | 20.544 | 0.0840 | 0.0055 | 82.16 |
| 13 | 18.427 | 7.033 | 12.837 | 0.0532 | 0.0089 | 126.07 |
| 14 | 22.273 | 10.144 | 19.133 | 0.0887 | 0.0060 | 88.66 |
| 15 | 21.747 | 9.714 | 10.612 | 0.0485 | 0.0108 | 151.58 |
Sensitivity analysis based on these results showed that the body thickness (\(DS\_D1\)) had the most significant influence on all output responses: mass, deformation, and stress. It exhibited a strong positive correlation with mass and a negative correlation with deformation and stress. This is intuitive, as a thicker body increases stiffness and reduces stress but adds weight. The Pareto front analysis from the response surface optimization identified the optimal design point as the one minimizing mass while satisfying all constraints. This corresponded to Run 15, with dimensions: \(DS\_D1 = 10.612 \, \text{mm}\), \(DS\_D2 = 9.714 \, \text{mm}\), \(DS\_D3 = 21.747 \, \text{mm}\). For practical manufacturing, these dimensions were rounded to \(DS\_D1 = 10.6 \, \text{mm}\), \(DS\_D2 = 9.7 \, \text{mm}\), \(DS\_D3 = 21.7 \, \text{mm}\). A final static analysis of this optimized geometry was conducted. The comparison between the initial and optimized designs is critical for assessing the success of the precision investment casting manipulator component optimization:
| Performance Metric | Initial Design | Optimized Design | Improvement / Change |
|---|---|---|---|
| Mass (kg) | 0.0928 | 0.0484 | Reduction of 47.8% |
| Maximum Deformation (mm) | 0.0057 | 0.0108 | Increase, but within limit (≤0.2 mm) |
| Maximum Stress (MPa) | 53.94 | 151.77 | Increase, but within constraint (≤180 MPa) |
The optimization successfully reduced the gripper mass by nearly 48%, achieving a significant lightweighting goal. While the maximum deformation and stress increased, they remained well within the prescribed safety limits, ensuring the structural reliability required for the demanding cycles of precision investment casting shell production.
Beyond static strength, the dynamic characteristics of the manipulator gripper are vital for stable operation, especially when integrated into a high-speed robotic system. Excessive vibration can lead to positioning errors, accelerated wear, or even resonance failure. Modal analysis was performed to determine the natural frequencies and mode shapes of the gripper, both before and after optimization. The analysis solves the eigenvalue problem derived from the undamped free vibration equation:
$$(K – \omega_i^2 M) \phi_i = 0$$
Where \(K\) is the stiffness matrix, \(M\) is the mass matrix, \(\omega_i\) is the i-th natural angular frequency (\(\omega_i = 2 \pi f_i\)), and \(\phi_i\) is the corresponding mode shape vector. The first six natural frequencies for the initial and optimized gripper designs were extracted:
| Mode Order | Initial Natural Frequency (Hz) | Optimized Natural Frequency (Hz) |
|---|---|---|
| 1 | 894.33 | 989.58 |
| 2 | 1051.90 | 1636.70 |
| 3 | 1278.00 | 1797.40 |
| 4 | 1278.90 | 1945.90 |
| 5 | 1424.20 | 2268.60 |
| 6 | 1450.30 | 2271.10 |
The mode shapes typically involved bending and twisting of the gripper arms and body. A crucial aspect is to ensure that the fundamental (first) natural frequency is sufficiently higher than any dominant excitation frequencies in the operating environment. In this application, the primary excitation could come from the servo motors driving the manipulator’s joints. Assuming a maximum motor speed of \(n = 3500 \, \text{rpm}\) with a fluctuation of \(\delta = 100 \, \text{rpm}\), the potential excitation frequency \(f_{ex}\) can be estimated, considering harmonics, as:
$$f_{ex} = \frac{n \pm \delta}{60} \times \text{(Number of Poles/Other factors)} \approx \frac{3500 \pm 100}{60} \times 2 \approx 113.3 \pm 3.3 \, \text{Hz}$$
This yields a range up to approximately 116.6 Hz. The first natural frequency of the optimized gripper is 989.58 Hz, which is over eight times higher than this estimated maximum excitation frequency. This substantial margin effectively avoids the risk of resonance, ensuring stable and precise operation during the rapid movements required in an automated precision investment casting cell. Interestingly, the optimization not only reduced mass but also increased the natural frequencies. This counter-intuitive result is possible because the material removal was strategically performed from regions contributing little to stiffness but significant to mass, thereby increasing the specific stiffness (stiffness-to-mass ratio) of the component.
The entire design and optimization process underscores the effectiveness of integrating computer-aided design, finite element analysis, and advanced optimization techniques in developing high-performance components for industrial automation. For precision investment casting, where process consistency and quality are paramount, the reliability of automation equipment is non-negotiable. The optimized shell-making manipulator gripper meets all functional requirements: it provides a secure grip with a mechanically efficient linkage, possesses adequate static strength with a safety factor above 2.5 even after optimization, and exhibits excellent dynamic characteristics with high natural frequencies far from operational excitations. The weight reduction of over 40% contributes to lower inertial loads on the robot arm, potentially allowing for smaller actuators, faster acceleration, and reduced energy consumption—all valuable benefits for a cost-effective and efficient precision investment casting production line. Future work could explore the use of alternative lightweight materials like aluminum alloys or composites for further mass reduction, or investigate the fatigue life under cyclic loading typical of continuous shell-making operations. Additionally, prototyping and experimental validation of the gripper’s performance in an actual precision investment casting environment would be a logical next step to confirm the simulation results and refine the design for real-world robustness.