An Evaluation Method for In Situ Form Errors of Large-Scale Mandrels Based on Biobjective Optimization Including Frame Dimension and Harmonic Suppression

Large-scale mandrels are indispensable core tooling equipment in high-end manufacturing for aerospace, energy equipment, and other critical fields. However, existing evaluation methods for in situ form errors of mandrels mainly emphasize harmonic suppression in error separation, while neglecting the influence of measurement frame geometric dimensions, which limits their applicability in compact measurement systems. To address this limitation, an evaluation method for in situ form errors of mandrels based on a synchronous three-probe configuration is proposed. A coupled analytical model integrating measurement frame geometric dimensions and error transmission characteristics is established, revealing the intrinsic relationship between probe angular distribution, harmonic suppression, and structural constraints. On this basis, a biobjective optimization strategy is developed to simultaneously minimize harmonic suppression effects and the measurement frame size, enabling a systematic tradeoff between measurement accuracy and structural compactness. Simulation results indicate that the proposed method achieves harmonic suppression performance comparable to existing approaches, while reducing the measurement frame geometric dimension by 52.9 % and maintaining the roundness measurement error below 0.1 %. Experimental results under different clamping conditions indicate that consistent mandrel profiles are obtained, with roundness values of 91.2 and 93.9μ m, respectively, despite significant variations in raw probe signals. These results illustrate that the proposed method suppresses spindle motion error effectively while enabling a more compact measurement system. The proposed method provides a new perspective for integrating structural compactness into error separation and can be readily extended to in situ form errors evaluation of other large-scale rotational components.