I am currently trying to introduce a constant load increase in a wind/diesel simulation.  Do you know how this can be done?  We have to assume an annual linear load increase for the load profile of 8 % over 15 years.

HOMER assumes a constant load throughout the project lifetime, so it does not let you explicitly model load growth. 

For now, you can model load growth in an approximate way by entering load data for some point near the middle of the project lifetime.  For example, if the project lifetime is 20 years, I sometimes scale up the load data to what I expect it to be in year 10.

If you are willing to do more work than that, you can model load growth more accurately by doing multiple simulations with increasing load size, essentially modeling each year of the project lifetime separately.  You would have to assemble the results into a single spreadsheet that showed key outputs like fuel consumption and operating cost for each year of the project lifetime.  Then you would have to do your own summation to calculate total net present cost.

It would be hard work to simulate a lot of system configurations in that way, but if your preliminary modeling has identified two or three possible configurations, you could model each of them in this more involved way to choose between them.