Opex appears to scale quadratically with N_WEC rather than linearly
Furthermore, the linear scaling was calibrated to be exact at N=1, but it could be recalibrated at N=100 to improve the nominal N=100 estimate.
Even though capex scales nicely with N_WEC, the amount of scaling coming from structures is wrong at N=100 (material cost of 4.26 vs 2ish in CBS sheet 1.4), since again this was calibrated at N=1.