Browse Prior Art Database

Single Transducer Noise Detection and Reduction Approach for Mud Pulse Telemetry

IP.com Disclosure Number: IPCOM000243010D
Publication Date: 2015-Sep-08
Document File: 6 page(s) / 672K

Publishing Venue

The IP.com Prior Art Database

Abstract

An automatic noise detection and reduction approach is described that can operate using one pressure transducer. Noise is detected in the frequency domain and removed using a time domain approach. Telemetry signal is transformed into frequency domain to detect the noise frequencies using a peak detection approach. Frequency tracking is applied to increase robustness and preciseness of the approach. The detected frequencies are removed afterwards using an IIR filter.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 49% of the total text.

Page 01 of 6

Single Transducer Noise Detection and Reduction Approach for Mud Pulse

Telemetry

An automatic noise detection and reduction approach is described that can operate using one pressure transducer. Noise is detected in the frequency domain and removed using a time domain approach. Telemetry signal is transformed into frequency domain to detect the noise frequencies using a peak detection approach. Frequency tracking is applied to increase robustness and preciseness of the approach. The detected frequencies are removed afterwards using an IIR filter.

Mud pulse telemetry signals suffer from two main noise sources; drilling noise and pump noise. There is nothing specific for reducing the drilling noise except High Pass Filter (HPF); however there are several approaches to reduce the pump noise effect.

These approaches often need extra hardware beyond only one pressure transducer; either extra pump stroke counters to get the pump signature or two pressure transducers. Accordingly, our focus will be implementing a notch filter, that can be automatically selected, as can be done using only one pressure transducer.

Pump noise is characterized by its harmonic behavior as shown in the following figure:

[SAM1]

Our development is to implement a notch filter, which can automatically select the frequencies to be notched.

Detailed Description:



Page 02 of 6

In order to reduce the effort for the field engineer by automating the system. As a step towards reducing the needed hardware, as a one sensor noise reduction approach.

The approach used for applying the notch filter:

1. Transform a time domain partition (e.g. 30s) to frequency domain (FFT). 2. Accumulate/average the amplitude spectrum across multiple time sections.

3. Define a threshold to differentiate between signal and noise peaks.

4. Apply harmonic frequency check to verify the detected noise peaks.

5. Apply notch filters at the verified frequencies.

6. Frequency tracking approach can be added to increase the algorithm robustness.

[SAM2]

The first step in the approach is the windowing in time domain (partitioning). Windowing is used to minimize edge effects that result in spectral leakage in the FFT spectrum. By using Window Functions correctly, the spectral resolution of your frequency-domain result will increase.


Page 03 of 6

Selecting the optimum window to use depends on the nature of the signal's spectrum and the particular harmonic to be analyzed. In our case we will consider the rectangular and the Hanning (Hann) windows.

After partitioning in time domain; each window is transformed into frequency domain using the Fast Fourier Transform (FFT). Then accumulation of absolute values is done per frequency bin over several partitions.

The number of partitions to be accumulated indicates the delay to update the notch frequency value. The current suggestion is to do time domain partitioning of ݐseconds then give the first update after n partitions, and keep accumulating windows till...