Primer for the ASOS Software Version 3.10 Ice Free Wind Sensor Quality Control Algorithm - PDF

Please download to get full document.

View again

of 12
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information Report
Category:

Ancient Egypt

Published:

Views: 5 | Pages: 12

Extension: PDF | Download: 0

Share
Related documents
Description
Primer for the AO oftware Version 3.10 Ice Free Wind ensor Quality Control Algorithm Prepared by: Chet chmitt, Field ystems Operations Center/Observing ystems Branch (OP22) Phone: xt 214 -Mail:
Transcript
Primer for the AO oftware Version 3.10 Ice Free Wind ensor Quality Control Algorithm Prepared by: Chet chmitt, Field ystems Operations Center/Observing ystems Branch (OP22) Phone: xt 214 -Mail: Bird activity and occasional ice build up on the Ice Free Wind ensor (IFW) has generated erroneous wind observations throughout the AO Network: unrealistic high gusts/peak winds, bogus variable wind direction reports and inaccurate average wind speed and direction reports. Bird activity has resulted in numerous extraneous priority 1 trouble tickets. An algorithmic solution has been incorporated into AO Version 3.10 software which quality controls wind data from the IFW. The algorithm has been implemented in such a way so that a priority 1 trouble ticket is only generated when a hardware problem is detected. This prevents the generation of extraneous priority 1 trouble tickets. The IFW QC Algorithm operates on the data received directly from the sensor: 1) very second, the wind direction and speed are sampled by the IFW. 2) very second, a running average of the most recent 3 seconds of data is computed by the IFW, producing the 3 second peak 3) very 5 seconds, the average of the most recent 5 seconds of data is computed by the IFW, producing the 5 second average. The highest 3 second peak is determined by the IFW and is stored as the 3 second peak. Discrete 5 second average Discrete 5 second average Discrete 5 second average Discrete 5 second average Time (seconds) W5 = 6.2 W3 = : Running 3 second average (i.e. 3 sec. peak) : Indicates 5 second discrete average to which the corresponding 3 sec peak is assigned W5, W3 (as determined above) along with diagnostic information is received every 5 seconds from the sensor: QC Algorithm: How it works The QC Algorithm evaluates each 5-second sample from the IFW against 9 criteria. amples failing to meet any of the 9 criteria are flagged. The QC Algorithm also assesses the pattern of flagged data to determine if the data stream itself is suspect. Flagged samples are recorded by AO and bracketed in the high resolution wind data archive and are T used in any of the AO wind algorithms. amples are also given a numeric error code corresponding to where the sample did not meet quality control. 5-second samples from IFW Flag samples when: 1. P/F flag from the sensor is F 2. ignal quality is less than (W3 peak - W5 avg ) W5 avg = 12 AND (WD5 avg WD3 peak ) W5 avg = 12 AND W3 peak (2.5 * W2Min) 6. W5 avg 12 AND W3 peak W2Min = 6 AND W3 peak 6 AND W3 peak (2.5 * W2Min) 8. W5 avg 165 OR W3 peak WT5 5 OR WT3 3 The Quality Control Algorithm also makes the following checks on the data stream: If 7 or more of the preceding 24 samples have been flagged (75% rule), all subsequent samples will be flagged until there are 18 consecutive samples which meet the QC criteria. T: When the 2-minute average wind speed exceeds 35 knots, only checks 1, 8 and 9 are made. valuate ample Against QC criteria Does the sample meet ALL criteria? High Level QC Algorithm Logic Mark the sample as suspect. DO T pass the sample into the AO Wind algorithms, bracket in the high resolution wind data archive. Has there been 18 consecutive samples that have met all QC criteria since the 75% rule was last violated? Pass the sample into the AO Wind algorithms. Detailed QC Algorithm Logic Begin IFW QC Algorithm Poll the wind sensor. Did the sensor respond? Obtain five second sample from the sensor. Obtain the two minute average wind speed from AO Two minute average wind speed (W2min) 35 knots? N O Data Quality Flag provided by sensor fail? WT3 = 3 seconds AND WT5 = 5 seconds 165 W3 = 0 AND 165 W5 = 0 W3 W5 -1 W5 12 AND W3 30 N O W5 = 12 AND WD5 WD3 30 N O ignal Quality = 79 et QC flag to fail Is W2min available from AO? W5 = 12 AND W3 (2.5 * W2min) N O W2min = 6 AND W3 6 AND W3 (2.5 * W2min) DO T pass sample to wind algorithms Did 7 or more out of the last 24 samples contain a QC flag set to fail? et Data tream Quality flag to fail Pass sample to wind algorithms N O Is the Data tream Quality flag set to fail? Is the QC flag for any of the last 18 samples set to fail? et Data tream Quality Flag to pass QC Algorithm: Impact on AO Maintenance The maintenance flag ($) and the corresponding priority 1 AOMC trouble ticket can only be set when a hardware problem is detected (sensor, communications, etc.). The maintenance flag ($) will T be set solely due to a wind data quality error, as purely environmental factors (such as roosting birds) are not issues that are correctable by the electronics staff. xtraneous IFW-related maintenance flags ($) and priority 1 trouble tickets are eliminated with software V3.10. QC Algorithm: Impact on AO Observations The IFW QC Algorithm results in a median loss of 0.33% of the good 5- second wind data. The IFW QC Algorithm results in the removal of 97% of corrupted 5- second wind data. QC Algorithm: Impact of Missing Data on AO Observations When the QC Algorithm removes six or less 5-second samples in a 2 minute period, the removal of those samples will be transparent to the end user: a QC ed wind report is available. When the QC Algorithm removes more than six 5-second samples, the wind report will go missing and will take a minimum of 4 ½ minutes to return. In order to remove 97% of the corrupted 5-second samples, the IFW QC Algorithm must remove some good data as well. The median amount of good data removed by the algorithm is 0.33%, which will be representative of the vast majority of the AO Network. There will be some sites that will experience good data loss of 1% or more, and other sites where good data loss will be near 0. ites with unorthodox siting constraints, sites in mountain passes and in dessert areas tend to have above median good data rejection rates. The majority of good data rejected are ordinary samples. However, the algorithm can wrongfully reject a piece of good data that is of operational significance. Case in point: the highest peak wind observed during a severe thunderstorm in Topeka, K on April 11, 2008 would have been rejected by the algorithm. T: Data rejected by the algorithm is not lost, it is retained for 14 hours in the high resolution archived. QC Algorithm: Impact of Bogus Data Removal on AO Observations Corrupt samples result in erroneous peak winds and gusts that can exceed 100 knots on a calm day. Corrupt samples result in erroneous variable wind reports, squall reports and wind shift reports. Corrupt samples result in unrepresentative wind speed and direction reports which can negatively affect surface analyses, forecaster awareness, forecast verification, the climate record and airport operations. Corrupt wind samples are T meteorological in nature, but are a byproduct of how the sensor interprets the interplay between the acoustical waves it uses to measure wind velocity and an entity in the sample volume (birds, ice, insects, etc.). The IFW QC Algorithm removes 97% of the corrupted 5-second samples, nearly eliminating bogus wind reports from the AO Network. QC Algorithm: Cost vs. Benefit Cost: The QC Algorithm causes 0.33% of the good wind data to be rejected from the AO wind algorithms. This can result in a missing wind report or the under reporting of a peak wind on rare occasions. Benefit: 97% reduction in bogus peak wind reports, bogus gusts, false variable wind reports, unrepresentative wind observations and false Q reports limination of extraneous priority 1 AOMC trouble tickets The tradeoff: 0.33% of AO 5-second wind data is not processed through the wind algorithms in exchange for a 97% reduction in bogus wind observations and the elimination of environmentally induced wind-related extraneous priority 1 trouble tickets. Anonymous Quote: how me a perfect algorithm, and I will show you an algorithm that does nothing.
Recommended
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x