MachinePlan-10K: A BRep-Based Synthetic Dataset for Machining Process Planning with Tool-Aware Feature Sampling

We present a dataset developed for machining process planning. The dataset was generated using Siemens Designcenter NX and contains 10,000 unique, validated parts (as BRep) with associated process plans, in-process workpiece (IPW) geometries, tool information, machining time and tool paths. The dataset focuses on planar machining operations under 2.5-axis CNC machining and is intended for research and development in machining process planning. All dimensions are in metric units (mm), and only features that can be machined using tools available in the NX CAM tool library are included. The automated data generation pipeline ensures consistent and accurate labeling, minimizing bias and discrepancies arising from manual annotation. Similarly, leveraging the NX knowledge base ensures that process planning rules are applied consistently in a deterministic way, eliminating variability from subjective judgments.