Page 192 - Fluid Power Engineering
P. 192

164   Chapter Eight



        Uncertainty Analysis
              Wind resource assessment is not complete without uncertainty anal-
              ysis. The reason is wind resource assessment provides information
              about wind speeds, which is transformed into average annual energy
              production (AEP), which is transformed into average revenue of the
              project. Uncertainty in wind resource is translated into uncertainty in
              AEP, which is translated into uncertainty in revenue. Any financial
              analysis of a wind project must consider not just the average revenue,
              but also the uncertainty in revenue. The most common method of
              incorporating uncertainty is to compute P50, P84, P95, and P99 esti-
              mates of revenue. Consider an example that has mean annual revenue
              of $2000 and standard deviation of $300. Table 8-6 contains the val-
              ues for PN. Most financial decisions are made based on P95 or higher
              estimates of revenue. Therefore, a financier is keenly interested in the
              estimate of σ and the methodology used to compute it. A bankable
              wind assessment should contain revenue estimates for various levels
              of PN along with a methodology for estimation of uncertainty.
                 Typical sources of uncertainty in energy production estimates in-
              clude:
                  1. Sensors. This was discussed in detail in Chapter 6. The un-
                    certainty includes sensor’s inaccuracy, calibration inaccu-
                    racy, and inaccuracy because of mounting and setup, and
                    other factors (like, overspeeding for cup anemometers, large
                    volume-based measurement for Sonic Detection and Ranging
                    (SODAR))
                  2. Shear model. This was also discussed in Chapter 6. This is the
                    uncertainty because of extrapolation from sensor heights to
                    hub height.
                  3. The spatial distribution model. This pertains to the uncertainty
                    introduced when wind data is translated from measured lo-
                    cations to the individual wind turbine locations.

                                PN Revenue
               PN Formula        Estimate  Meaning of PN
               P50 = μ            $2000    Revenue of at least $2000 will be
                                           realized with 50% certainty
               P84 = μ − σ        $1700    Revenue of at least $1700 will be
                                           realized with 84% certainty
               P95 = μ − 1.645 σ  $1506    Revenue of at least $1506 will be
                                           realized with 95% certainty
               P99 = μ − 2.326 σ  $1302    Revenue of at least $1302 will be
                                           realized with 99% certainty

              TABLE 8-6 Uncertainty Expressed as PN
   187   188   189   190   191   192   193   194   195   196   197