基于独立成分分析的复杂曲面加工误差分离

Machining error separation of complex surfaces based on independent component analysis

研究基于独立成分分析的复杂曲面加工误差分离技术,旨在将复杂曲面加工误差按照不同的误差源进行分离,克服当前复杂曲面加工误差分离时无法分离相近尺度误差的问题.对各种误差源所导致的加工误差进行建模,然后阐述独立成分分析的原理和步骤;通过对混合后的信号进行分离,判断原始的4种信号与ICA分离出的4种信号之间相关性的大小,验证ICA用于解决分离混合信号问题的可行性;最后以实际加工中心加工的复杂曲面为例进行实例验证,结果表明:应用ICA可以有效地将复杂曲面相近尺度的加工误差进行分离,有利于进一步对各误差源进行判定.

This study was based on the independent component analysis(ICA)of complex surfaces machining error separation technique.The aim was to separate the machining errors of a complex surface according to the different error sources,overcome the problem that the surface machining error separation cannot be separated from the similar scale error.Firstly,the machining errors caused by various error sources were modeled.Then the principles and steps of independent component analysis were described.Through the simulation of signal separation,the correlation size of simulation signals and ICA isolated signals was determined and then the feasibility of ICA was proved.Finally the actual complex surface was taken as an example for verification.The results showed that the complex surface machining errors of similar scale were effectively separated by ICA.It was helpful to determine the various error sources further.