Emerson Rosemount 700XA Reference Manual page 26

Overview
June 2022
being recognized as true peaks by the program. It is therefore important to choose a
peak width that corresponds to the narrowest peak in the group under consideration.
1.4.8 Peak detection
For normal area or peak height concentration evaluation, the determination of a peak's
start point and end point is automatic.
The manual determination of start and end points is used only for area calculations in
Forced Integration mode. Automatic determination of peak onset or start is initiated
whenever Integrate Inhibit is turned off. Analysis is started in a region of signal quiescence
and stability, such that the signal level and activity can be considered as baseline values.
Note
The controller assembly software assumes that a region of signal quiescence and stability
will exist.
Having initiated a peak search by turning Integrate Inhibit off, the controller assembly
performs a point by point examination of the signal slope. This is achieved by using a
digital slope detection filter, a combination low pass filter and differentiator. The output is
continually compared to a user-defined system constant called Slope Sensitivity. A default
value of 8 is assumed if no entry is made. Lower values make peak onset detection more
sensitive, and higher values make detection less sensitive. Higher values (20 to 100) would
be appropriate for noisy signals, (e.g., high amplifier gain).
Onset is defined where the detector output exceeds the baseline constant, but peak
termination is defined where the detector output is less than the same constant.
Sequences of fused peaks are also automatically handled. This is done by testing each
termination point to see if the region immediately following it satisfies the criteria of a
baseline. A baseline region must have a slope detector value less than the magnitude of
the baseline constant for a number of sequential points. When a baseline region is found,
this terminates a sequence of peaks.
A zero reference line for peak height and area determination is established by extending a
line from the point of the onset of the peak sequence to the point of the termination. The
values of these two points are found by averaging the four integrated points just prior to
the onset point and just after the termination points, respectively.
The zero reference line will, in general, be non-horizontal, and thus compensates for any
linear drift in the system from the time the peak sequence starts until it ends.
In a single peak situation, peak area is the area of the component peak between the curve
and the zero reference line. The peak height is the distance from the zero reference line to
the maximum point on the component curve. The value and location of the maximum
point is determined from quadratic interpolation through the three highest points at the
peak of the discrete value curve stored in the controller assembly.
For fused peak sequences, this interpolation technique is used both for peaks, as well as
valleys (minimum points). In the latter case, lines are dropped from the interpolated valley
points to the zero reference line to partition the fused peak areas into individual peaks.
The use of quadratic interpolation improves both area and height calculation accuracy and
eliminates the effects of variations in the integration factor on these calculations.
26
Reference Manual
2-3-9000-744
Emerson.com/Rosemount