基于数据驱动的惯性测试设备倾角回转误差检测

打开文本图片集

中图分类号：TH703 文献标志码：A 文章编号：2095-2945（2025）31-0027-04

Abstract：Adata-drivenmachineleamingmethodisproposedtomeettherequirementofdetectingthespindleinclination rotationerrofadual-axisturntable.Adual-channellaserdisplacementsensormountedonthehousingisusedtocolectyaw dataoftemainshaftintwoorthogonaldirectionsinrealtime，andtheaverageyawvalueduringonecycleofmotionis obtainedasaninputfeature;theinclinationrotationerorvalueobtainedbytraditionaldetectionmethodsisusedasalabelto buildadataset.Theyawdataandtargetinclinationrotationerorarenormalizedaccordingtotheknownphysicalupperand lowerbounds.Buildamulti-layeredperceptronmodel，usetheAdamoptimizerduring thetrainingprocesssetEpochto100， and avoid overfitting with an early stop strategy. Experimental results show that on the training set， MSE≈8×10-6 ， MAE≈2.2×10-3 and R2≈0.9999 ； on the test set， MSE≈1.1×10-5 ， MAE≈2.6×10-3 and R2≈0.9998 . It isverified that the proposed method can predict the rotation error of the spindle tilt angle without the need for autocollimator.

Keywords：inertialtestequpment;incliationrotationeror;datadriving;atificialneuralnetwork;laserdisplacementsensor dual-axis turntable

双轴转台主轴的倾角回转误差直接影响测试精度和系统可靠性。（剩余5600字）