The transformation from a finished part design to a blank model is critical for manufacturing efficiency, especially for casting parts. This process requires sophisticated feature recognition and conversion techniques to bridge CAD and CAM domains. Current methodologies primarily focus on feature recognition and generation, with significant research dating back to foundational work by Kyprianou in 1980. Early approaches, such as Arbab’s feature decomposition method, utilized Boolean subtraction operations on raw material to derive part geometry. Subsequent research by institutions like Cranfield理工学院 and Berlin Technical University advanced feature classification and boundary representation models. However, these methods often neglect manufacturing constraints specific to casting parts, particularly regarding machinability and geometric relationships between features.
To address this gap, we present an integrated approach for reconstructing blank models of casting parts. The methodology classifies manufacturing features into two categories:
| Feature Type | Definition | Characteristics |
|---|---|---|
| Face Machining Features | Thin features with constant cross-section | Formed in single operation (e.g., planes, small holes) |
| Volume Machining Features | 3D material removal features | Require multi-surface machining (e.g., pockets, slots) |
The reconstruction process follows four critical stages:
1. Predefined Feature Filtering: We establish manufacturing rules to identify features unsuitable for casting. For holes in casting parts, minimum castable diameters vary by production scale and method:
| Casting Method | Mass Production | Single Production |
|---|---|---|
| Sand Casting | 30 | 50 |
| Metal Mold Casting | 10-20 | – |
| Pressure Casting | 5-10 | – |
The filtering operation is defined as:
$$S_v = S_p \cap S_d$$
$$L(P) = P \cup S_v$$
where \(S_p\) = part features, \(S_d\) = predefined features, \(S_v\) = filtered features, and \(L(P)\) = filtered part model.
2. Post-Casting Surface Reconstruction: Convexity testing identifies machined surfaces. For each face \(F_k\) with specified tolerances/surface finish:
$$ \text{Convexity}(F_k) = \begin{cases}
\text{Machined} & \text{if } \forall E_i \in F_k, \text{ } E_i\text{ is convex} \\
\text{Cast} & \text{otherwise}
\end{cases} $$
Manufacturing knowledge rules determine machining allowances \(\Delta S\):
$$ \Delta S = \sum_{i=1}^{n} \Delta_i $$
where \(\Delta_i\) = allowance per operation. The surface is then offset by \(\Delta S\) along its normal to generate the blank surface, creating elevated model \(R(P)\).
3. Intersecting Feature Processing: Non-convex surfaces indicate volume features. We apply half-space modeling where adjacent half-spaces \(H(F_i)\) and \(H(G_i)\) combine via manufacturing-driven rules:
$$ \text{Merge}(H_a, H_b) = \begin{cases}
H_a \cap H_b & \text{if convex edge} \\
H_a \cup H_b & \text{if concave edge}
\end{cases} $$
The resulting volume \(V_o\) is unified with \(R(P)\) to form the blank model.
4. System Implementation: Our system architecture integrates these modules:
- CAD Interface: Imports casting part designs
- Feature Processor: Executes filtering and reconstruction
- Knowledge Base: Stores manufacturing rules and allowances
- CAM Output: Exports machining features
This methodology significantly advances blank modeling for casting parts by automating allowance calculation and feature recognition. Current limitations include incomplete handling of fillets, drafts, and parting surfaces. Future work requires standardized feature representation formats and enhanced design-blank associativity for full CAD/CAM integration. As casting parts grow in geometric complexity, robust blank generation becomes increasingly vital for manufacturing efficiency.