PSD-based parallelism measurement method of guideway and uncertainty evaluation

Abstract This study proposes a method for measuring guide rail parallelism based on position-sensitive detectors (PSDs), establishing a comprehensive uncertainty assessment framework. Parallelism errors are calculated using the public benchmark method and least squares fitting, with uncertainty components quantified via a dissemination of the value of quantity model. The reliability of the assessment is verified using the Monte Carlo method, yielding a 95% confidence interval of 10.0–14.0 μm. The innovative introduction of a conjugate prior Bayesian framework enables dynamic updating of repeatability uncertainty, reducing the expanded uncertainty by 10% (from 2.0 μm to 1.8 μm, k = 2). Comparative laser interferometer experiments demonstrate that while the interferometer offers higher precision (0.8 μm), the PSD system maintains 1.8 μm accuracy at a 33% lower cost (from ¥300 000 to ¥200 000). It supports 8 h of operation without an external power supply, offering superior portability and industrial adaptability.