Fatigue Design of Wind Turbines - Wind Turbine Technology: Fundamental Concepts of Wind Turbine Engineering - page 692

above the band indicate fatigue load problems that are probably cost-effective to solve by de-
sign changes. Responses below the band are a reasonable goal for improved HAWTs, since
these would be expected to experience fatigue loads that are smaller than those in most of the
test cases of the data set. It should be carefully noted that empirical equations of this type are
not to be relied upon to form the sole basis for a complete fatigue load analysis.
A useful application of Equations (12-7a) to (12-7d) is in the verification of structural
dynamic computer models, such as those discussed in Chapter 11. As mentioned before,
computer codes of this type are becoming more and more reliable for the prediction of me-
dian cyclic loads. The above empirical equations are a convenient method for calibrating a
computer model of the structure, to verify that predicted median fatigue loads are consistent
with HAWT field test data. Once this is done, high-percentile load predictions from the
computer model can be compared to predictions made with the empirical equations in order
to “tune” the model and its wind turbulence inputs.

Design Load Trend Charts

One of the most important uses of empirical equations is to evaluate the effects of po-
tential design changes through parametric studies during the conceptual design phase of a
project. These are also referred to as trade-off studies , because technical benefits must be
balanced or traded off against any impacts on system costs and schedules. Design trend
charts are convenient tools for displaying the relative effects of configuration and operational
changes, and the general procedure for preparing them is as follows:

-- Define a baseline configuration and calculate baseline fatigue loads.
-- Vary one parameter as an independent variable while holding the remaining
parameters constant at their baseline values or scaling them in proportion to
the independent variable.
-- Calculate loads for the changed configurations and normalize ( i.e., divide)
each of them by the appropriate baseline load.
-- Graph the results as load ratios vs. the independent parameter, for desired
probabilities of exceedance.

This process will now be illustrated by sample calculations in which Equations (12-7a) to
(12-7d) are used to calculate the relative changes in both fatigue loads and stresses that are
caused by increasing and decreasing the rotor diameter [Spera 1994].

Baseline Configuration and Loads

A baseline configuration that represents the mid-range of the rotor, tower, and site data
in Tables 12-4 and 12-5 can be defined in terms of the following parameters and factors:

D = 46.0 m U 0 = 7.9 m/s

P R = 410 kW U 2 = 14.3 m/s

d 3 = 0 deg M g = 456 kN-m

c t = 0.80 m q = 5 deg

Z = 660 m s = 0.066

H = 38.0 m a 0 = 0.40

N = 30.0 rpm a = 0.232

w c = 152 cpm Tower type = shell, upwind of rotor