For the param file are not the same of different enzyme libraries. How to combine two replilcates of two enzyme libraries into pairs. And how to do normalization between different enzyme libraries and test the difference by diffHic. 

This is not supported in diffHic, and there are no plans to do so. Rigorous integration of Hi-C data generated with different enzymes is challenging, because the interaction-level and genomic biases (dependent on cut site frequency, fragment length, etc.) will not cancel out between libraries. This means that the biases would have to be modelled explicitly, which requires strong assumptions to describe the process by which the biases are generated.