Actually the effect of dilation angle can be related to its influence on the mobilization of the soil friction angle. Based on the research made by Bolton during 1980s, there is an empirical relationship between the mobilized friction angle and the dilation angle. Of course this is not the only effect but can be considered as one of the important ones. Please also note that the failure mechanism may be subjected to change by including the angle of dilation.
if you carry out simple direct shear test on a dense sand applying different normal stress, you will obtain typical shear stress vs displacement curves. Please note you will see higher values of dilation at lower normal stress. At a very high normal stress you are likely not to get any dilation effect as soil particles can't roll over each other. Therefore, if the applied bearing stress on sandy foundation soil is very high, it would be prudent to use constant volume angle of friction rather than peak angle of friction.
another question please, when use same friction angle and the dilation during the loading, the curve of load vs displacement it's linear stable, however, when friction angle higher than dilatation (non-associated flow rule), curve its unstable.
i want understand role this dilation angle in the behavior of soil ( how the volume of soil is made during loading )?
I cannot fully understand the "stable" and "unstable" responses you mentioned before. There is nothing about the associativity and non-associativity when the stability (I mean stability in response not stability against failure) is investigated. I recommend you read the Drucker's "stablity postulate". You can find basic materials on this topic in the second chapter of "Limit Analysis in Soil Mechanics, Chen and Liu (1990)".