MEO/WEO (%) Expected performance* (η) is the ratio between Measured Energy Output (MEO) from PCV tests, and estimated energy output from OEM’s warranted power curve (WEO) η = MEO/WEO (%) Loss factor (LF) is the estimated turbine underperformance resulting from historical industry-wide or OEM-specific performance data. LF varies depending on the consultancy. 108,0 107,0 106,0 105,0 104,0 103,0 102,0 101,0 100,0 99,0 98,0 97,0 96,0 95,0 94,0 93,0 92,0 StdDev IE 20 IE 13 IE 21 IE 16 IE 6 IE 18 IE 10 IE 7 IE 1 IE 11 IE 15 IE 19 IE 8 IE 2 IE 17 IE 5 IE 12 IE 9 2.5 2.3 2.5 3.8 2.7 2.1 1.4 2.3 1.4 2.1 1.7 4.0 0.7 3.1 2.1 2.1 2.0 1.8 Figure 1 Results of expected performance of Vestas turbines calculated by 18 different external consultancies. Each box plot represents the median value of the expected performance, as well as the range of the results for each consultancy. Vestas has observed a variation of 3.3 percent points in the standard deviation of the expected performance4, resulting from the different conditions at the site for the measurement and calculation methods. This confirms the lack of consistency in the methodology applied as well as the negative impact of including the climatic conditions in the PCV tests*. Energy production Assessment Capturing the value of accurate power curve predictions For project planning, energy production estimates are an essential part of a business case. Wind park-specific energy production evaluations are based on climatic conditions and the warranted turbine power curves that are provided by the turbine manufacturer. These warranted power curves, which are a result of the OEM’s production estimates, are based on the design choices of technology and performance modelling, which are also validated by measurement campaigns. To maximise certainty in estimated wind park energy production, external consultants are often engaged to analyse warranted power curves provided by the manufacturer and conclude on a project’s expected performance. External consultants’ evaluations are typically built from measurement campaigns on actual wind turbine performance, commonly known as Power Curve Verification (PCV) tests , which are then utilized for the estimation method of the specific site. Current lack of transparency and standardisation in calculating expected performance Recent research on energy production assessments has uncovered highly underappreciated risks for end users e.g. developers, financing institutions and investors1 2: 1. High variance observed in default expected performance discounts (loss factor) Without sufficient measurement data, external consultants apply a default discount associated to the expected performance, known as loss factor, to all turbine manufacturer’s estimated power output. While this number may appear precise, it is actually the result of a standardised average of data within a wide range (figure 1). Regardless of a particular turbine’s performance, this arbitrary reduction is applied to business case calculations. In fact, DNV GL estimates that turbines across the industry underperform up to 4% as compared to the power curve provided by the manufacturer3. Furthermore, UL AWST, recently optimised their own power curve loss factor by reducing it 0.3 percent points1. The industry need to standardise and increase transparency of the methodology and data used for calculating the loss factor is raised by both DNV GL and UL AWST, who have assessed more than one third of the global wind parks’ expected performance in 2019. The fact that DNV GL has acknowledged variations in their calculations and UL AWST has adjusted their own power curve loss factor only reinforces the importance of this industry need. 2. The competitiveness of new technology is impacted by default expected performance discounts (loss factor) As the pace of innovation accelerates, turbines are increasingly sold in earlier phases of development. Due to insufficient or non existent operational data from the actual wind turbine, a default loss factor is applied whenever a new turbine model is released, whereby impacting the true competitiveness of the product. Independent consultants have concluded that a particular turbine’s expected performance is most correlated with both design choices and manufacturer—not the default loss factor1 . This correlation is so strong that UL AWST has recommended the concept of performance families to give more accurate calculations when insufficient measurement data is available1 . In addition, further development and standardisation of power curve prediction models that incorporate both historical data and modelling assumptions would mean more accurate calculation of the loss factor for each turbine.
Download PDF file