sets reflect the technology market prices at a specificmoment in time. TheETYS15cost snapshot could bedominated by an offer-driven-market resulting in highprices, which might not be valid as a long-term averageneeded for the overall results.Since the strongly simplified cost models ofImperialCollegeandTorbaghanare not suitable for estimatingthecostofback-to-backprojects,theiroverall10
performance cannot be determined (mathematicallyleading to an infinite error).8. Discussion8.1. Limitations of the methodologyThere are a few simplifications and assumptionslimiting the validity of this analysis, as described below.8.1.1. Parameter set conversionTheparametersetsobtainedfromtheliteraturehave been converted into the structure of the linearinvestment cost model proposed in Section 3. In manycases, this conversion causes the loss of information.For instance,Bk0has been estimated to zero if it couldnot be determined by the available parameter set data.Nevertheless, a conversion of the cost parameter datato fit the proposed linear cost model is unavoidable,since otherwise hardly any comparison would have beenpossible.8.1.2. Line type cost ratiosA simplified cost ratio between submarine cablesand underground cables as well as overhead lines hasbeen used for the evaluation. Assuming different ratioscould have minor effects on the parameter evaluationfor projects with large underground cable or overheadline sections.However, the simplified cost ratios arenecessary to consider the contractual cost taken frompress releases.8.1.3. Investment cost markupA blanket markup on the contractual cost figureshas been used when evaluating cost estimates againstreference project cost.As previously mentioned, thecost markup is necessary to consider the contractual costtaken from press releases.8.1.4. Evolution of costThe analysis does not include the cost changes overtime, e.g.learning curves or inflation correction forthe cost parameter sets.Given the limited overallreference data availability, learning curve effects havebeen excluded as they are, at least for the moment,impossible to quantify. The inflation correction for thecost parameter sets could have been included. However,as learning curve and inflation effects presumablyinfluence the costs in opposite directions, the accuracyof the results is expected to be better when neglectingboth, opposed to neglecting only one of them. It wastherefore chosen not to correct the cost parameter setsfor inflation.8.1.5. Biased validationDue to the fact that the limited available referencedata used for the validation might or might not bethe same as the one used for creating some of thecost parameter sets, a possible methodological flawcould arise.If they, by any chance, coincide whilecomparing the cost parameter performance against thereference data, the observed amount of deviations couldbe lower, providing a false sense of reliability for thecost parameter set in question.
You've reached the end of your free preview.
Want to read all 17 pages?
- Fall '10
- Electric power transmission, Cost overrun, VSC, VSC HVDC