APPENDIX A - Technical data and Methodology:

The beaufort data was for 13 locations, for 12 months plus year total, giving hours for 7 speed bands 0 through 11 metre/second at 2 metres above ground level - 91 data pairs per location, 1183 data pairs total. Data was input manually using dbase3+, and data entry checking (using checksums) was performed within the program as detailed below.

The anemograph data was similar, for 4 locations, approx 364 data pairs total; heights were all 10 metres - these were corrected; some windspeeds were in mph not knots - all were converted to metre/second.

The cca data was 13 data points (monthly plus annual) for 59 locations - 767 data points total.

All beaufort and anemograph data was converted to cumulative hours and to metre/second at 2 metres height.

Total energy for the 221 location-periods (17 locations * 12 months + 1 annual) was calculated as a straight integral of v^3.dt, using velocity steps of 0.1 metre/second (100 steps on average per location-period).

Windrun was also calculated as integral of cca speed wrt time, using 3 different possible cut-in speeds for the Casella Mark II cca's - 1.2 m/s, 1.7 m/s and 2.24 m/s.

The maximum impulse WEC extractable energy for each location-period was calculated by stepping through a range of cut-in windspeeds using iterative techniques; this was stored, as was the optimum cut-in windspeed vciopt corresponding to this Emax. The formula used was similar to that described in (15) above - the characteristic of an impulse type WEC; the integral was taken as vci^2 * vblade, where vblade/v tends to 1.00 as v/vci tends to infinity.

Then for each location period the vci values corresponding to 80% and 90% of Emax, both above and below vciopt were calculated using iterative techniques.

The program was written in Clipper 87 and took approximately 7 hours to run through on a 386DX with 12 msec hard drive. Graphs as below were done in Harvard Graphics, using ascii data import from data exported by the program to text files. These graphs for space and cost reasons are not included with this report as standard, BUT CAN BE SUPPLIED on floppy disc and/or hard copy if requested; they are of course central to much of the value of this report, and certainly central to proving that the report's findings are valid.

For the 13 beaufort sites and separately for the 4 anemograph sites the vcca simulated were compared with the vcca actuals. For the beaufort data it was found that the cca simulated and actual were a very good correlation using a cca cut-in windspeed of 2.24 m/s (as stated by Parkes); for the anemograph data there was no fit whatsoever using any of the 3 possible cut-in windspeeds - thus it was decided to discard the anemograph data as inaccurate for all purposes and applications.

The total energy, maximum impulse energy and vciopt were now graphed against the vcca simulated (all these 4 values were simulated), from the beaufort data only, i.e. discarding the anemograph data. This gave very good fits on the 2 energies, and less good but very useable on vciopt. The formulae as per (4a), (4b) and (5) above were obtained from these graphs.

The 80% and 90% vci's were now compared with the vciopt's - the 0.67, 0.78, 1.26 and 1.38 ratios were established as good fits as per (6) above.

The cca data for the 59 sites were processed to give Emax and vciopt for the 12 months plus annual average; also a recommended vcibest, which depended on seasonal variability, and the resultant values of Ebest extractable from that year-round cut-in speed. The data was presented in sequence of descending annual wind energy and in sequence of worst month wind energy (see appendix 2 below).

The seasonal energy yield data was also presented in context of normal rainfall occurrence, so as to give some indication as to complementary or otherwise energy requirement for water pumping.