Near Infrared Optical Sensing Protocols for Wet End Pulp Slurry Verification
Inline NIR transflectance sensors stabilize wet end pulp slurry consistency within +/- 0.3 percent, cutting fiber giveaway and dryer steam consumption.
Savitzky Golay Filter
Signal Smoothing
Digital data processing algorithms remove high frequency noise from continuous measurements by fitting successive subsets of adjacent data points with a low degree polynomial through linear least squares. In paper manufacturing, a savitzky golay filter corrects fluctuations in basis weight profiles captured by beta gauges across moving webs. Retaining underlying signal trends without shifting peak positions distinguishes this mathematical operation from standard moving averages.
Operational tolerances on high speed converting lines demand precise boundary conditions to prevent artificial distortion of narrow grammage peaks.
Noise Reduction
Industrial scanners generate millions of data points per minute where electrical interference obscures true variations in substrate caliper. Applying weighted moving averages introduces a phase lag that shifts the apparent position of defects along the running web. Convolution coefficients assigned to local windows preserve the height and width of sudden caliper spikes while suppressing random measurement scatter.
Machine operators rely on clean spectral profiles to adjust slice lip actuators automatically without chasing false alarms caused by sensor jitter.
Profile Stability
Automatic control loops depend on uncorrupted feedback signals to maintain uniform cross direction thickness profiles during winding. Unfiltered data triggers unnecessary actuator corrections that create cyclic variations in roll hardness and subsequent web breaks in printing presses. Mathematical smoothing prevents mechanical wear on pneumatic positioning equipment by eliminating jittery adjustment commands.
Stable transverse profiles ensure consistent ink transfer across coated art papers during high speed offset printing.