Michael W. King, Phillips Petroleum Company, Idaho Falls,
Idaho
T he purpose of the program is to fit the best sequences of
parabolas to a given 400 point data curve in order to remove
extraneous noise; rather than rely on a single 400 point
parabola least squares fit to approximate a given data curve.
Approximately 400 individual parabolas are computed as
follows.
LESQ29
Data values 1 through 29 are subjected to a second order
Least Squares fit. The median point of the resulting
parabola (point #15) is then substituted for the original data
value #15.*
A second parabola is then computed using data values 2
through 30. The median point of this parabola (point #16)
is then substituted for point #16 of the original data curve.
This procedure is repeated until all data values have been
replaced (except for the first and last 14 points which are
excluded by the mechanics of the operation).
*See B. J. Power, R. N. Hagen, S. O. Johnson, "SPORT,
A System for Processing Reactor Transient Data on the IBM-
7040 Computer, " pp. 4-8, AEC Research and Development
Report (IDO-17078), Available from: The Clearinghouse for
Federal Scientific and Technical Information, National Bureau
of Standards, U.S. Department of Commerce, Springfield,
Virginia.
LESQ11
Process identical to LESQ29 except that an 11 rather than
a 29 point smooth interval is used. First point replaced is
point #6, and only the first and last 5 points are excluded
from smoothing.
LESQ11 will preserve higher frequency data than LESQ29
for a given data curve with constant time between data
points.
Minimum Hardware: | 4K PDP-5 or PDP-8,
Teletypewriter (plotter, DECtape
optional)
|
Other Programs Needed: | Floating Point Interpretive Package
(Digital 8-5-S) and appropriate
data handling routines.
|
Storage Requirement: | LESQ11: 400-564; 700-716
LESQ29: 400-564; 700-751
|
Execution Time: | (PDP-5) LESQ11: 1 minute.
LESQ29: 2.5 minutes.
|
Restrictions: | Positive integer data <37778; time
between data points constant.
|
Catalog:
| November 1969
|