The Royal Netherlands Meteorological Institute (KNMI) is planning the installation and/or replacement of cloud ceilometers and present weather sensors at oil production platforms in the North Sea. KNMI operates the new Vaisala-Impulsphysik LD40 ceilometer and the Vaisala FD12P present weather sensor in the national measurement network at land station, but has no experience with these sensors on the North Sea. However, KNMI experience at stations near the coast show that often the regular maintenance interval of 6 months is not sufficient for the FD12P due to contamination. In order to determine whether these 2 new sensors are suitable for use on the North Sea and to get an estimate of the contamination and required maintenance interval a test has been performed. For that purpose a LD40 ceilometer and a FD12P present weather sensor were installed on Meetpost Noordwijk (MPN), a research platform near the Dutch coast, form March 13 to November 18, 2002 (Figure 1 shows an overview of the set-up). During that period the standard output of the sensors, that is used in the operational set-up, is stored on a PC every minute as well as status information and additional information that is a measure of the contamination. This additional information consisted of the window backscatter (WBS) of the LD40 and the receiver (RBS) and transmitter backscatter (TBS) of the FD12P as well as the signal of the precipitation detector (DRD) of the FD12P (Table 1 shows an example of the archived data).
The measurements of the LD40 show that the sensor works correctly during its 8-month stay on MPN without maintenance. The window backscatter remains below the 50% alarm limit during the entire period (Figure 5 shows the development of the WBS). Only on August 23 the sensor reports 2 alarms due to contamination, but these occurred during suspicious circumstances when the WBS suddenly jumped to a higher value (Appendix B contains daily plots of LD40 data showing the observed cloud hits as well as the WBS). The LD40 that was used in the test had an old laser module that resulted in faulty cloud detections above 10,000 feet, but this did not influence the test. Also the cloud hits at 25 feet that occur on some days without fog seem not related to contamination. The measurements also show that as a result of precipitation the contamination of the window of the LD40 generally decreases. Furthermore it can be noted that the sensor (including the plastic hood) survived and operated correctly during the storm on October 27 with wind gusts up to 38m/s.
The measurements of the FD12P at MPN show that after about 6 months the sensor is seriously contaminated and in urgent need of maintenance (Figure 12 shows the development of the RBS and TBS). The sensor already reports warnings due to contamination of the transmitter only a couple of days after installation; after about a month and a half the sensor reports such warnings almost continuously; after 6 months the first errors are reported and the last months such error occur frequently, although the situation improves a little at the end. The transmitter is pointed towards the southwest and the contamination increases particularly during situations with both precipitation and a southwesterly wind. The contamination of the receiver looking to the northeast is much less. Still, warnings occur regularly after about 5 months. Again the situation improves a little in the last weeks of the test. Analysis of the frequency distribution of the measured visibility shows that the measurements are affected by contamination (see Figure 14). The contamination of the FD12P at Meetpost Noordwijk resulted in higher visibility values. This effect becomes noticeable when the reported contamination of the transmitter is about halfway between the corresponding warning and alarm limits (see Figure 15). The variation of the measured receiver backscatter suddenly increases on November 3 for no apparent reason and remains high. Low receiver backscatter values are reported since then with values that are even below the reference value for a clean receiver. The reference value for window contamination of the receiver is therefore no more valid and the reported backscatter is probably no good measure anymore for the contamination. The changes in the backscatter of the receiver not only occur during precipitation, but also on days without precipitation and/or high wind speeds. Furthermore, precipitation can lead to an increase or a decrease of contamination or no significant change at all. The precipitation detector shows no sign or contamination since the maximum, i.e. dry, values are nearly the same over the period of the test. Low values do only occur on days with precipitation or with high relative humidity.
Wiel Wauben. Duurtest LD40 en FD12P op Meetpost Noordwijk
KNMI number: TR-256, Year: 2003, Pages: 62