We currently create the frequency vector using the fundamental (f1) and the number of frequencies (nfreq). However, this means that you can often wind up with a lot of low frequencies where no energy is present (see, e.g., omega < 1 below). One could reduce the pseudo-spectral problem size by cutting out these lower frequencies.

https://github.com/SNL-WaterPower/WecOptTool/blob/9d8f4ef76c86cedf8ed997d85e1698f2bebc15b7/wecopttool/core.py#L1264-L1291