A turbine takes momentum out of the wind, so the air behind it is slower and more turbulent, and the next turbine downwind gets less power. The Jensen (Park) model treats the wake as a cone that widens linearly with distance; inside it the speed is reduced by a top-hat deficit. Overlapping wakes add in quadrature, and the power follows a cubic curve between cut-in and rated speed.
single wake: dU/U = (1 - sqrt(1 - Ct)) / (1 + 2 k x / D)^2 wake radius r = D/2 + k x
partial rotor: dU_ij = (A_overlap / A_rotor) x single-wake deficit
several wakes: dU_j = sqrt( sum_i dU_ij^2 ) U_j = U_free,j (1 - dU_j)
power curve: P = Pr (U^3 - Uin^3) / (Ur^3 - Uin^3) for Uin < U < Ur, P = Pr up to Uout (3 / 12 / 25 m/s, Pr = 3 MW)
thrust: Ct = 0.8 up to rated speed, then falling as (Ur/U)^3
terrain: U_free(x,y) = U (1 + beta h(x,y)/h_max) (hills accelerate the flow)
yield: P_mean = sum over 12 sectors f_s (3 sub-angles) and speed bins w(U) of P_array, AEP = 8766 h x P_mean
- Row versus staggered - aligned columns are the worst when the wind blows along a row; staggering moves wakes between turbines in the main direction but loses the benefit when the wind turns, which is why the rose-averaged number can favour either.
- Spacing is cheap to compute, costly to build - efficiency rises with spacing roughly as the wake deficit falls with (1 + 2 k x/D)^2; the chart shows which directions cost most.
- Terrain - a ridge crest accelerates the wind, so a ridge line can out-yield a valley layout even with more wake loss in cross winds. Here the speed-up is a simple height-proportional factor, not a flow solver.
- Same model in both pages - one shared physics module; the checks compare it with the closed-form single wake, a grid integration of the overlap, a far-apart limit, the Weibull bins and a Monte Carlo draw.
- Honest simplifications - top-hat Jensen wake (real wakes are Gaussian and meander), no turbulence or atmospheric stability, no wake recovery over terrain, the same Weibull curve in every sector, no electrical, availability or blockage losses, fixed rotor diameter 100 m and hub height 90 m, made-up terrain. Teaching model, not a yield assessment.