Research On Model Identification Of Ultra-precision Motion Stage

Lithography is a high-precision manufacturing equipment of electromechanicalintegration, involving precision machinery, precision control, precision measurementand other subjects. It is limited by the development of optical properties, electricalproperties, measurement resolution, machining capabilities, control performance, etc.The performance of lithography can be improved by the design of controllers that canensure the high speed and high precision.The core devices of lithography is wafer stage and reticle stage, which is composedof coarse and fine motion stage with gas-lubricated bearings. Permanent magnet linearmotor and voice coil motor drive the coarse and fine motion stage respectively, so as toachieve six degrees of freedom motion. For optimizing the mechanical structure andvalidating of advanced control methods, it is necessary to identify the ultra-precisionmotion platform model with resonance characteristics.This dissertation analyzes the structure of precision motion platform and buildsmathematical model of linear motors and voice coil motors, thereby SDOF macro-microcoupling theoretical models and6-DOF mechanical model are established, which canreflect the combined effect of multiple motors motion characteristics. Theoretical modelis the basis of model identification.Data acquisition and data pre-processing are the key links of system identification.Scanning frequencies experiment has been done on the system. Data interpolationalgorithm and filtering method are designed for processing the gross error appearing inthe data. Correlation analysis method is applied to calculate the output signal amplitudeand the phase difference relating to the input signal. The frequency characteristics ofcoarse motion stage and fine motion stage can be reflected by Bode diagrams drawnrespectively.The unknown parameters of the SDOF macro-micro coupling theoretical model areidentified by adaptive real-coded genetic algorithm. Considering the great disturbancegenerated by mechanical resonance, the mechanical resonance model is built, and theparameters of the model are identified based on the hierarchical identification principle.Then the notch filter that can be used to suppress the mechanical resonance is designedto improve the dynamic response of the system.Validity of the precision motion platform model has been verified. The model willbring great convenience in designing and debugging of lithograph control system.