Agriculture Reference
In-Depth Information
Table 12.2
Errors of yield sensing systems for combine harvesters at different throughputs in the
2000/2001 test bench studies
Sensing system (in bold),
Meter
(in italics),
Manufacturer (in brackets)
Relative calibration
error
in % (average)
Standard deviation
of relative error
in %
Light-barrier, system 2,
CERES 2,
(RDS)
−0.57
5.50
Light-barrier, system 2,
QUANTIMETER,
(CLAAS)
−2.71
1.72
Light-barrier, system 2,
PRO SERIES 2000,
(RDS)
−3.89
5.54
Radiation, system 3,
FLOWCONTROL,
(MASSEY FERG.)
−1.64
3.02
Force, system 4,
FIELDSTAR,
(DRONNINGB./AGCO)
−0.22
1.52
Force, system 5,
YM 2000
(AGLEADER),
LH 565
(LH AGRO)
−1.71
3.65
Force, system 6,
GREENSTAR,
(JOHN DEERE)
−2.89
2.81
Flat standing position; 10, 15, 20, 25 and 30 t h
−1
throughput; 5 repetitions/variant; n = 25/m;
reference mass/variant = 1 t; winter wheat. For system numbers see Fig.
12.2
(Demmel
2001
)
by all meters. The standard deviation (s) is the measure of the sensing accuracy.
It indicates the range of error, within which around 2/3 of all measurements lie.
Despite the different measuring principles, all sensing systems are characterised by
approximately equal ranges of error between ±3.5 and ±4 %.
In the
test bench studies
, the accuracy of the systems was determined under
identical, clearly defi ned conditions. Special consideration was given to the effect of
different throughput levels and of transverse and longitudinal tilt (Table
12.2
).
When the measuring accuracy of the various yield measuring systems was checked
in the test bench under fl at conditions at different throughputs, mean calibration
errors <3 % are obtained with two exceptions. Larger deviations (3-10 %) only occur
at lower throughputs (10 t h
−1
). This indicates that the calibration curves plotted in
the instruments are not optimally matched to low throughputs.
Across all throughputs, the standard deviations varied between 2 and 6 %
(Table
12.2
). At the individual throughput levels, the standard variations were only
between 0.5 and 3 % (not in Table
12.2
). Distinct differences between the sensing
systems do not exist.
However,
lateral and longitudinal tilts
of the combine harvesters at constant
throughputs (20 t h
−1
) exert a much greater infl uence upon the accuracy of the meters
(Table
12.3
).
The least deteriorating infl uence of tilt on the errors results when radiation is
used for the sensing (system 3). In order to compensate for the infl uence of tilt,
some volumetric measuring systems are equipped with one or two axle tilt sensors.
Nevertheless, the errors caused by lateral and longitudinal tilt cannot successfully
be compensated under all conditions. In this regard, the force measuring systems
occupy an intermediate position between radiometric and volumetric meters.
Taylor et al. (
2011
) collected data from 29 test plots for corn on farms in Kansas,
Alabama and Iowa, which were harvested using different yield measurement
