Trading days effect in natural gas and electricity series

[Barbora00]

Hi,
I am doing a project on GARCH models within the energy markets - gas, electricity, oil using daily data (1 month forward prices 2002-2013). Oil seems to be fine but in electricity and gas time series (residuals, squarred residuals respectively) there seems to be significant autocorrelation around the lags 20-22 apparently trading days effect (as there are around 22 trading days in one month). I have experience just with modelling the stock index FTSE, DAX where this has never happened. Could someone give me a feedback please, whether you have encountered such a situation? And maybe some papers which deal with that? The series were seasonally adjusted and the day-of-the-week was not significant, so no need to account for that. Thanks in advance!
Barbora