Page 118 - Fluid Power Engineering
P. 118

96   Chapter Six


                 The next data management is a recurring task that is performed
              daily or with a frequency that coincides with data transfer from tower
              to server. This validation of data is a significant task in data manage-
              ment. Typical sources of errors and validation issues with measured
              data include:

                    Wear and tear of sensors, booms and mounting, guy wires,

                    and met-tower
                    Icing of sensors

                    Lightning strike

                    Loose or faulty wiring

                 In validation check, three tests are performed on the data: 8

                    Range test: This test identifies values that are outside the val-

                    idation criteria. Table 6-4 contains a sample of the validation
                    criteria. The offset is the lowest value recorded by sensor when
                    the wind speed is below the sensor’s threshold. A reading of
                    wind speed above 25 m/s does not necessarily mean that the
                    data is suspect, but it requires a review.
                    Relational test. This test compares concurrent measured val-

                    ues of redundant sensors, and sensors at different heights.
                    If the differences do not meet the validation criteria, then a
                    review is required. Sample test criteria are in Table 6-5.
                    Trend test. This test examines temporal changes in measured

                    values. Sample test criteria are presented in Table 6-6.
              After the data issues have been identified, the next step in Fig. 6-13
              titled “Validate Data” is executed.


        Data Processing
              In order to perform statistical analysis of the measured data and to
              visualize the data, a variety of data processing methods are used.
              For illustration purposes, publicly available wind speed data from a
              National Renewable Energy Laboratory (NREL) project in Valentine,
              Nebraska, is used in this section. The data is available at heights of 10,
              25, and 40 m above the ground level, and for a period of 4 years.
                 For statistical analysis, the data is organized into bins. An exam-
              ple of two-dimensional binning of data is seen in Table 6-7. The two
              dimensions for the bins are wind speed and direction. For wind direc-
              tion, it is normal to use 16 bins with bin size of 22.5 ; some programs
                                                        ◦
                                      ◦
              use 12 bins with bin size of 30 . The wind direction bins are called sec-
              tors. Figure 6-14 shows the frequency plot of the North-North-West
   113   114   115   116   117   118   119   120   121   122   123