TABLE OF CONTENTS
3. DATA COLLECTION HISTORY
|
Meteorological Field Measurements at Potential and Actual Wind Turbine Sites | |
|
TABLE OF CONTENTS
| ||
|
3. DATA COLLECTION HISTORY
| ||
During the past four years when PNL conducted the measurement program at
the candidate sites, the following procedures were established for performing
maintenance, calibration, and data reduction by PNL's subcontractor. These
procedures were followed until June 1981 when PNL assumed all site
maintenance and calibration and data processing responsibilities. Procedures
now followed by PNL are described at the end of this section.
4.1 CALIBRATION AND MAINTENANCE
The subcontractor provided routine calibration and maintenance services,
as well as emergency maintenance services, at each site that received DOE
equipment. For example, when a utility representative reported a sensor or
logger failure, the subcontractor proceeded to the site within 72 hours to
correct the problem (in some instances, the problem could be identified and
corrective procedures implemented by discussing the situation with the
utility representative over the telephone).
The subcontractor also performed calibrations at each of the sites. Annually, calibrations included changing all sensors with precalibrated spares. Quarterly onsite calibrations of all electronics were performed. The precalibrated wind speed sensors were calibrated in a wind tunnel. (WSSI utilized the Colorado State University wind tunnel facilities at Fort Collins, Colorado; ESC used the NOAA Atmospheric Turbulence and Diffusion Laboratory (ATDL) wind tunnel at Oak Ridge, Tennessee; and E-S used the wind tunnel facili ties at MRI, Inc., Altadena, California). All calibrations were traceable to standards established by the National Bureau of Standards.
4.2 SITE MAINTENANCE,
DATA EDITING, AND REDUCTION
At the operational sites, a representative of the utility visited the tower site weekly and changed the cassette tape in the data logger every other week, except at turbine sites where cassettes were changed weekly. Before June 1981, the data-filled cassette was mailed to PNL's subcontractor.
The utility representative also visually inspected the tower and sensors, checked the performance of the logger, replaced the rechargeable batteries as needed, and filled out a site inspection form that was also mailed to the subcontractor or PNL. Completion of the site inspection form was an important part of both the maintenance and data editing portion of the program. On this form the utility representative was asked to note existing meteorological conditions occurring at the site as well as readings from the various data channels of the data logger. This information aided the subcontractor and PNL meteorologists in determining if a possible equipment failure had occurred or appeared likely and the approximate time during the month the problem occurred. The latter information was valuable in processing and editing the data from the cassette tape.
PNL's subcontractor utilized a computer program to convert the data on the cassette data tapes to engineering units. These data were then edited by a computer program to determine if
If any of these conditions occurred, an appropriate message was printed out along with the particular scan in question and a meteorologist reviewed this listing to determine if the data should be changed or coded as missing. Later, PNL requested the subcontractor to code as missing values the obvious bad data and to flag questionable data by a lettering scheme that was coded along with each data scan within the magnetic tape. Once the data were properly edited, they were put on a nine-track tape, a copy of which was forwarded to PNL for analysis.
The subcontractors were also required, as part of their quality assurance program, to perform a routine "dummy data analysis" on all computer programs that were involved in data processing. This procedure would document that the individual computer programs were performing as expected. This technique is commonly used in quality assurance procedures involving data collection at nuclear reactor facilities (Sandusky and Conley 1979).
4.3 DATA REPORTING
When the edited monthly nine-track data tape was received at PNL, it was put into a data analysis program, which provided summaries of pertinent wind characteristics at each candidate site. These summaries were incorporated into monthly candidate wind turbine site data reports, which were forwarded to NASA/Lewis, DOE, and each participating utility, thereby allowing an assessment of the wind energy potential at the sites. These reports contained information on
In addition, annual and five-year candidate wind turbine site data reports are published (Sandusky and Renne 1981a, 1981b; Sandusky et al. 1982), which are available from the National Technical Information Service (NTIS). These reports differ from the monthly data reports in that the two-minute data are first used to create hourly averages before the final analyses are made. This is required because of the vast amount of two-minute data collected at the sites over a period of a year. Data from the 17 original sites have also been summarized by Sandusky and Buck (1981).
4.4 QUALITY ASSURANCE
A Quality Assurance (QA) program, similar to that required by the United States Nuclear Regulatory Commission (USNRC), was established for the MVP. The objectives of the QA program were to ensure that all procedures for calibration, maintenance, and data reduction were carefully followed and results documented. The QA program required, for example, that PNL subcontractors verify that the engineering units created from the cassette data tapes were accurate conversions of the sensor output values as recorded on the tapes. A key element of the QA program was the filing of Nonconformance Reports, which documented the apparent time and type of any component failure and the steps taken to remedy the situation.
PNL's subcontractors designated a QA officer who ensured that all QA procedures were strictly followed. The QA officer was not directly associated with the calibration, maintenance, and data reduction activities and therefore retained a more independent, unbiased perspective of the program.
4.5 CURRENT PROGRAM AND DATA BASE DEVELOPMENT
In June of 1981 all site maintenance and calibration procedures and data processing functions were taken over by PNL. This was done in anticipation of a phase-out of the candidate site program. The calibration and maintenance activities were assumed by PNL technicians; however, the utility personnel who visited the sites periodically also contributed significantly to routine maintenance activities. For example, since the instrument elevator systems allowed the sensors to be brought down to ground level for routine change-out, and PNL supplied spare data loggers to each utility, it was simple for utility personnel to replace and return any damaged equipment to PNL for repairs. This also greatly reduced down time. Although data recovery rates have suffered somewhat since the 72-hour emergency maintenance procedure was discontinued (rates that had been upwards of 90% dropped to below 80% in some cases), these procedures have permitted conservation of program funds without drastic losses in usable data.
Perhaps the biggest advantage in this revised program management format has been in processing and reporting the data. The current data processing procedure, which has been in place at PNL since June 1981, bypasses the need to translate the cassette tapes onto a nine-track tape. Instead, as the cassettes are received at PNL from the participating utilities, they are read directly into a computer system and stored on a disk. The editing and processing are performed routinely as one step; when a monthly file of data has been assembled, a monthly data analysis report can be produced. This procedure has greatly increased processing efficiency, decreased the turnaround time for producing individual monthly reports, and has allowed the creation of a comprehensive candidate site data base. This data base has proved invaluable for performing special analyses and for responding to numerous data requests.
The data editing procedure was streamlined so that data were edited, coded as missing if required, and converted to engineering units by one computer program. This procedure, which had been suggested by the PNL subcontractors, was required because of the large volume of data to be processed once all sites were operational. As the data were processed, any values changed to a missing code were listed as part of the program. These listings were reviewed by a PNL meteorologist to determine if any "good" data were being incorrectly changed to a missing code. Generally, the above condition only occurred when abnormal time-of-day or Julian day data were recorded on the cassette tape. In this case the PNL meteorologist, using the text editor of the computer, could locate the abnormal data scan and delete that scan or change the incorrect value. Since the original data, stored as a binary file, remained on disk, the editing and engineering conversion procedure was repeated and subsequent outputs from the program were again reviewed by the meteorologist. Thus, the editing procedure for some individual tapes required several iterative review and reprocessing steps.
Another innovation that has taken place in the program since June of 1981 has been the installation of remote data transfer systems at some of the sites. These systems store up to 32 hours of hourly data collected at the site, then transmit the data over existing phone lines to a printer or to the computer system at PNL. A total of nine units were installed at sites where phone lines were readily available by early 1982, including the MOD-DA turbine installation sites identified in the next section.(a) With the implementation of an automatic dial-up system in the computer, it is possible to maintain a daily file of current data in the computer. This allows for more efficient identification of any problems with the sensors at the sites and allows the data to be used in other wind energy studies, such as wind forecasting studies, where rapid access to the data is required.
(a) Except at the Culebra, Puerto Rico, site
5. TURBINE INSTALLATION SITES
Return to RReDC home page ( http://rredc.nrel)