Faulty DSM time series

[khelfen]

The DSM time series appear implausible at the moment.

• [x] p_max_pu and p_min_pu are not clipped at 0
• [ ] p_max_pu and p_min_pu are sometimes larger than p_set (after a discussion with @KathiEsterl and @ClaraBuettner this seems to be justifiable)
• [x] The wrong p_set time series is saved