band-reject

Band-reject filtering of downhole drilling vibration sensor data removes the discrete-frequency interference that contaminates accelerometer, strain gauge, and shock sensor recordings in measurement-while-drilling and rotary steerable system tools operating in Western Canada Sedimentary Basin horizontal wells, where the dominant noise sources are the fundamental rotation frequency of the drill string (typically 2 to 6 Hz for 120 to 360 rpm top drive speeds) and its harmonics at integer multiples, the bit bounce frequency excited by PDC cutter engagement with hard Montney siltstone or Duvernay limestone stringers (typically 5 to 20 Hz), and the 60 Hz power-line frequency that couples into the surface data acquisition electronics through grounding loops between the rig electrical system and the MWD surface unit, each of which occupies a narrow discrete frequency that can be eliminated by a notch (band-reject) filter tuned to that specific frequency without degrading the broadband vibration signal used to characterize shock severity, whirl, stick-slip, and lateral vibration modes that damage the BHA. The band-reject filter differs from the band-pass filter used in MWD telemetry decoding in that it removes specific narrow frequency bands while passing everything else, rather than selecting a narrow passband and rejecting everything outside it; in drilling vibration analysis the vibration signal of interest is broadband (0 to 200 Hz for shock severity), so any filter that attenuates a broad frequency range would destroy the diagnostic content of the record, making the notch-style band-reject filter the only acceptable noise removal approach. In WCSB horizontal drilling with top drives rotating at 180 to 240 rpm (3.0 to 4.0 Hz fundamental), the rotary harmonics at 3, 6, 9, 12, and 15 Hz corrupt the vibration spectrum at exactly the frequencies where bit whirl (3 to 8 Hz), lateral shock (5 to 20 Hz), and torsional stick-slip oscillation (0.5 to 3 Hz) signatures are most diagnostic, requiring a comb notch filter that applies a set of narrow notches at the fundamental and the first four harmonics of the measured rotation frequency simultaneously, with each notch having a 3 dB bandwidth of 0.2 to 0.5 Hz and a stopband attenuation of 40 to 60 dB to suppress the rotary harmonic spikes that otherwise appear as false resonances in the vibration spectrum displayed to the driller on the surface vibration monitoring unit. Understanding band-reject filter design for drilling vibration applications (notch bandwidth selection, attenuation depth, comb filter implementation for multiple harmonics, real-time rotation frequency tracking that keeps the notch centered on the actual drill string RPM as the driller adjusts top drive speed), the trade-off between notch bandwidth and the risk of inadvertently suppressing genuine vibration signatures near the rotary harmonic, and the 60 Hz power-line notch design for surface data acquisition systems gives WCSB drilling engineers, BHA designers, and vibration monitoring specialists the signal processing tools to correctly identify downhole vibration modes that cause PDC bit damage, MWD tool failures, and drillstring fatigue in WCSB tight reservoir horizontal programs.

• Comb notch filter design for rotary harmonic suppression in WCSB MWD vibration monitoring: A comb notch filter for WCSB horizontal drilling vibration analysis applies N simultaneous narrow-band notches at frequencies f0, 2f0, 3f0, ... Nf0, where f0 is the current drill string rotation frequency in Hz. For a WCSB Montney lateral being drilled at 210 rpm (f0 = 3.5 Hz), a 5-notch comb filter attenuates 3.5, 7.0, 10.5, 14.0, and 17.5 Hz; each notch has a 3 dB bandwidth of 0.3 Hz and stopband attenuation of 50 dB, achieved with a 4th-order IIR notch filter at each frequency. The comb filter update rate must track the actual rotation frequency to within 0.1 Hz to prevent the notch from missing the harmonic spike when the driller adjusts top drive speed; WCSB vibration monitoring systems update the comb filter coefficients from the measured RPM input (from the top drive encoder or the MWD rotation rate channel) every 2 seconds, preventing filter drift that would allow the harmonic spike to migrate outside the notch bandwidth and reappear in the cleaned spectrum.
• 60 Hz power-line notch filter for WCSB MWD surface data acquisition system interference: The 60 Hz power-line frequency couples into WCSB MWD surface data acquisition systems through shared grounding paths between the rig generator panel and the MWD surface unit chassis, appearing as a discrete 60 Hz spike in standpipe pressure and vibration sensor recordings that can be mistaken for a genuine high-frequency formation or tool vibration signature. A single-frequency IIR notch filter at 60.0 Hz with a 3 dB bandwidth of 0.5 Hz (59.75 to 60.25 Hz) and 60 dB stopband attenuation eliminates the power-line interference with negligible effect on the broadband drilling vibration spectrum below 50 Hz. In WCSB operations where the rig generator frequency drifts by plus or minus 0.3 Hz under variable load (common for smaller 1,000 to 1,500 kVA diesel generators), the notch filter must be implemented as an adaptive notch that estimates the actual power-line frequency from the recorded spectrum and centers the notch accordingly, rather than a fixed 60 Hz notch that misses the drifted fundamental and provides less than 10 dB rejection.
• Bit bounce notch filter selection for WCSB PDC drilling in hard carbonate stringers: PDC bit bounce in hard formations produces a characteristic axial vibration at the bit-bounce frequency, approximately equal to the P-wave transit time resonance of the BHA length divided by two (typically 8 to 20 Hz for WCSB BHA lengths of 100 to 200 m). In WCSB Duvernay horizontal laterals that cross tight carbonate stringers and chert beds, bit bounce at 11 to 14 Hz coincides with the 3rd or 4th harmonic of the drill string rotation at 200 to 250 rpm, making it difficult to distinguish genuine bit bounce from rotary harmonic noise. The diagnostic approach is to measure the vibration spectrum during a brief rotation-without-bit-weight interval (keeping the BHA rotating at the same RPM but lifting WOB to zero), then compare the spectrum to the drilling-with-WOB spectrum; frequencies that appear only during drilling with WOB and disappear when WOB is removed are bit-bounce candidates rather than pure rotary harmonics, and a band-reject notch is applied only to the identified rotary harmonic components, preserving the bit bounce signature for severity assessment.
• Stick-slip torsional vibration frequency and the conflict with band-reject filter placement in WCSB drilling: Torsional stick-slip in WCSB Montney horizontal BHAs occurs at the torsional resonance frequency of the drill string, typically 0.3 to 2.0 Hz for 4,000 to 8,000 m string lengths, which directly overlaps the fundamental rotation frequency (3 to 6 Hz) only when severe stick-slip causes the drill string RPM to oscillate between near-zero and twice the average RPM at the stick-slip resonance frequency. If the comb notch filter is centered on the average RPM but the drill string is experiencing stick-slip with RPM variation of plus or minus 50% around the mean, the rotary harmonic notch may suppress the stick-slip signature by coincidentally filtering the frequency band where the stick-slip oscillation appears. WCSB vibration monitoring systems that use dynamic notch filters (tracking instantaneous RPM rather than average RPM) avoid this problem by shifting the notch with the instantaneous rotation rate, ensuring the notch follows the harmonic spike but does not suppress the torsional oscillation frequency that lies between harmonic positions.
• Band-reject filter validation using synthetic vibration injection in WCSB BHA qualification testing: Before deploying a new BHA design with integrated vibration monitoring in a WCSB Montney horizontal program, drilling engineers validate the band-reject filter performance by injecting a synthetic vibration signal of known frequency content into the accelerometer recording chain at the surface test bench and confirming that the filter correctly suppresses the injected rotary harmonics while passing broadband shock content. The synthetic injection test applies a multi-tone signal at 3.5, 7.0, 10.5, 14.0, and 17.5 Hz (simulating 210 rpm rotary harmonics) with amplitude equal to three times the expected formation vibration level, plus a broadband Gaussian noise component from 0 to 200 Hz with amplitude equal to the expected shock level; the filter output should show full suppression of the tonal components (greater than 40 dB reduction) and less than 1 dB alteration of the broadband noise spectrum between 18 and 200 Hz, confirming the notches are narrow enough not to degrade the shock and lateral vibration data used for real-time drilling optimization in WCSB tight reservoir programs.

Power-Line Notch Filter Resolving False Vibration Alarm on a WCSB Montney Well

A northeast British Columbia Montney horizontal well program experienced repeated high-frequency vibration alarms from the surface MWD monitoring unit during rotary drilling at 220 rpm through a 3,800 m lateral section. The vibration alarm threshold at 60 Hz indicated severe high-frequency lateral vibration that would normally prompt the driller to reduce WOB or RPM to protect the MWD tool. On inspection, the MWD field specialist found that the 60 Hz alarm corresponded to a spectral spike at exactly 60.0 Hz in the axial accelerometer record, with amplitude 15 times higher than the broadband background at adjacent frequencies. The spike was present at identical amplitude whether the driller applied WOB or lifted off bottom, confirming it was not formation-induced vibration. Investigation identified a grounding loop between the MWD surface unit chassis and the rig SCR house shared ground bar; the loop antenna had a coupling coefficient sufficient to inject 1.8 g equivalent peak acceleration at 60 Hz into the accelerometer recording channel. A hardware ground isolation transformer was installed on the MWD surface unit power input, eliminating the ground loop coupling; simultaneously, a software 60 Hz IIR notch filter with 0.5 Hz bandwidth and 55 dB attenuation was applied to all accelerometer channels as a permanent signal conditioning step, reducing the residual 60 Hz artifact to 0.02 g and eliminating the false alarm for the remainder of the drilling program.

Related Terms

Band-reject filter is the primary entry covering band-reject filter theory and seismic noise removal applications; this companion entry covers drilling vibration signal processing, where comb notch filters suppress rotary harmonic interference and 60 Hz power-line noise from downhole accelerometer data in WCSB MWD and RSS tools without degrading the broadband shock and vibration content used for BHA damage assessment. Drilling vibration is the broadband downhole signal that band-reject filters preserve while removing discrete harmonic interference; stick-slip, bit bounce, whirl, and lateral shock each produce characteristic frequency signatures in the 0 to 200 Hz band that drive BHA protection decisions in WCSB Montney and Duvernay horizontal programs. Measurement while drilling (MWD) tools contain the accelerometers and shock sensors whose recordings require band-reject filtering; MWD tool failures from undetected vibration are a leading cause of non-productive time in WCSB horizontal programs, making accurate vibration detection after harmonic noise removal critical. Stick-slip torsional vibration is the most damaging vibration mode in WCSB horizontal drilling and the one most at risk of being obscured or inadvertently suppressed by an overly broad rotary harmonic notch filter; the band-reject filter design must preserve torsional oscillation signatures at 0.3 to 2.0 Hz while rejecting the adjacent rotation fundamental. Rotary steerable system (RSS) tools in WCSB horizontal completions contain orientation sensors whose static and dynamic bias correction depends on clean low-frequency accelerometer data; 60 Hz and rotary harmonic contamination that leaks through an inadequate band-reject filter degrades survey accuracy and introduces systematic azimuth error in WCSB Montney lateral well placement.