There is no general approach to deal with your situation. The approaches that might be applicable will depend on the microarray platform, the experimental design, and the questions you want to answer. I would suggest consulting a local bioinformatics collaborator to help you with your questions--it will be much more effective than doing this electronically.

That being said, the Bioconductor project has tools for normalization and preprocessing as well as batch effect removal (if that is even applicable), meta-analysis, and differential expression.