I am working on a Microgrid planning and I discovered that the current existing generators of one of the communities are able to meet their demand within the given planning horizon of 20 years.(the is the result obtained from GAMS). When I implemted same model in Homer Pro,Homer replaced some gens even without using up the life of some other already existing gens. However, when I assume a very high replacement cost(higher than used in GAMS),homer uses the existing gens and didnt replace.I found this weird or there is something that I do not understand about how the two softwares works. I know most people here have worked with both softwares and could give an insight on the behaviours of these two softwares. Please I am really on a hot seat to answer this question and any help from any one here would be highly appreciated?
Generally, I believe that GAMS answers are more accurate. In return, Homer has a user friendly interface and it's a good mean to solve semi-complicated problems.
I'm not going deep into details, but let me say the final words first; LEAVE HOMER AND SOLVE YOUR PROBLEM WITH GAMS.
Direct answer is that even two different techniques solving the same problem might not end with accurate results having high correlation.
Homer is simple for implementation but difficult to optimize. GAMS is difficult to implement however once you have made the model then its easy to optimize. In Homer, cost of DG can be easily adjusted. But new generators after a certain planning horizon and load growth results in another optimization problem. On my suggestion, make one model on normal load and one model after including load growth after five years and optimize in both softwares. Give a comparison, if your results are close then fine, else no need to panic and above-mentioned reason is enough that you can find in various DG based planning papers.
The previous two responses pretty much cover the issues around choice of the tool. However, the details matter. HOMER is fundamentally not an optimization model - it uses heuristics that can lead to odd results. It is a low-cost user-friendly alternative to a full-blown optimization that one can implement using either GAMS or one of the many languages in that math programming family using really powerful optimization solvers (e.g., CPLEX and PLEXOS). There is of course the high end commercial planning software like PLEXOS that use the power of same optimization solvers, but are also user-friendly (except they are pricey!). So there are more choices and it is also a question of what kind of model you write in GAMS.
Having said all that I did not find results from HOMER as extreme as you have described. Perhaps, you should contact them and alert them to your findings to see how it might be resolved.
GAMS is one of the leading tool providers for the optimization industry and was the first software system to combine the language of mathematical algebra with traditional programming concepts in order to efficiently describe and solve optimization problems. While HOMER is a simulation model which will attempt to simulate a viable system for all possible combinations of the equipment that you wish to consider. The HOMER model greatly simplifies the task of designing hybrid renewable microgrids, whether remote or attached to a larger grid.
