Characterizing non-linear dynamic processes

Patent metadata

Field	Value
Publication number	US-2023154080-A1
Application number	US-202117995523-A
Country	US
Kind code	A1
Filing date	Apr 20, 2021
Priority date	Apr 20, 2020
Publication date	May 18, 2023
Grant date	—

Abstract

A non-linear dynamic process such as multiphase flow is characterized to determine a condition of the process. A sensor may obtain information about strain, vibration, flow rate, etc. during the process. That sensor data may be plotted, and at least one perceptual hash can be generated from the plot. The perceptual hash can be compared against a database storing plots or hashes for non-linear dynamic responses for different sensor types or processes. A best fit hash may be identified and used to determine what conditions are occurring in the field, or multiple near fits can define an envelope of experimental/reference conditions with similar non-linear dynamics to the field system. Multiple hash comparisons may be used to refine the envelope, determine a best fit, or determine what measurement would help discriminate the flow conditions.

First claim

What is claimed is:

1 . A method for identifying a condition of a non-linear dynamic system, comprising:
obtaining sensor data from the non-linear dynamic system; plotting the sensor data; creating one or more perceptual hashes from the plotted sensor data; comparing the one or more perceptual hashes to a set of reference perceptual hashes and thereby identifying one or more matching reference data sets; and providing information on the one or more matching reference data sets.

2 . The method of claim 1 , wherein plotting the sensor data includes generating a Poincaré plot.

3 . The method of claim 1 , wherein plotting the sensor data includes plotting the sensor data over a period of time that is less than 1 second, less than 2 seconds, less than 5 seconds, less than 10 seconds, less than 20 seconds, or less than 60 seconds.

4 . The method of claim 1 , wherein using the one or more sensors includes using a distributed acoustic sensor to obtain strain or vibration data from the non-linear dynamic system.

5 . The method of claim 1 , wherein creating the one or more perceptual hashes includes creating at least one of an average hash, a difference hash, a wavelet hash, or a DCT hash.

6 . The method of claim 5 , wherein creating the one or more perceptual hashes includes creating at least two different types of perceptual hashes.

7 . The method of claim 5 , wherein comparing the one or more perceptual hashes to the set of reference perceptual hashes includes comparing a composite of at least two perceptual hashes to a composite of at least two reference perceptual hashes.

8 . The method of claim 1 , wherein comparing the one or more perceptual hashes to the set of references perceptual hashes includes calculating a Hamming distance.

9 . The method of claim 1 , wherein creating the one or more perceptual hashes from the plotted sensor data includes defining a grid over the plot.

10 . The method of claim 9 , wherein the grid has a regular spacing.

11 . The method of claim 1 , wherein identifying one or more matching reference data sets includes identifying a plurality of matching reference data sets, the method further comprising:
for each matching reference data set, comparing one or more additional perceptual hashes from the plotted sensor data to one or more additional reference perceptual hashes to identify a higher likelihood matching reference data set.

12 . The method of claim 1 , wherein creating the one or more perceptual hashes from the plotted sensor data includes generating a binary number by replacing local values in a grid with a 1 or a 0 based on such local values being at least one of:
above an average; below an average positive; or negative.

13 . A method for determining a condition of a dynamic system, comprising:
receiving strain or vibration sensor data on a conduit having fluid flow therein; generating a plot for the received sensor data; generating at least one perceptual hash from the plot of the received sensor data; accessing a data store having a plurality of reference hashes corresponding to a plurality of reference data sets; and comparing the at least one perceptual hash to the plurality of reference hashes and, in response, identifying at least one matching reference data set of the plurality of reference data sets.

14 . The method of claim 13 , wherein identifying the at least one matching reference data set includes identifying a best matching reference data set, identifying a plurality of reference data sets having a best matching value, or identifying a plurality of reference data sets that exceed a threshold matching value.

15 . The method of claim 14 , wherein the threshold matching value is at least 0.8, 0.85, 0.9, 0.91, or 0.92.

16 . The method of claim 14 , further comprising:
defining an envelope of experimental conditions based on the plurality of reference data sets; and at least one of:
narrowing the envelope by comparing one or more additional perceptual hashes to the plurality of reference hashes; or
determining an additional measurement type that would provide discrimination narrowing the envelope.

17 . The method of claim 13 , further comprising at least one of:
raising an alert when the at least one perceptual hash does not match an expected set of hashes from one or more envelope reference data sets of the plurality of reference data sets; or using the identified at least one matching reference data set to calibrate the one or more sensors.

18 . The method of claim 13 , the one or more sensors including at least two sensors and the at least one perceptual hash includes a plurality of field perceptual hashes, the method further comprising:
evaluating whether the plurality of field perceptual hashes generated from data from the at least two sensors maintain an expected pattern.

19 . A dynamic system, comprising:
one or more fiber optic or acoustic sensors; one or more processors; and one or more computer-readable media including computer-executable instructions that, when executed by the one or more processors, causes the system to:
obtain sensor data from the one or more sensors;
plot the sensor data;
create one or more perceptual hashes from the plot of the sensor data; and
compare the one or more perceptual hashes to a set of reference perceptual hashes and, in response, determine a type of fluid flow within the dynamic system.

20 . The dynamic system of claim 19 , wherein creating the one or more perceptual hashes includes:
defining a grid for a 2-D non-dimensional plot; constructing a histogram of the 2-D non-dimensional plot on the grid; generating a value in each cell of the grid based on the histogram; and using at least one of values corresponding to cells in the grid, or differences between values corresponding to adjacent cells, in creating the one or more perceptual hashes.

Assignees

• Schlumberger Technology Corp

Inventors

• Williams Michael John
• Oddie Gary Martin

Classifications

• G06T11/26

  Drawing of charts or graphs · CPC title

• G01N29/4472

  Mathematical theories or simulation · CPC title

• G01F25/10

  of flowmeters · CPC title

• G01F1/74

  Devices for measuring flow of a fluid or flow of a fluent solid material in suspension in another fluid · CPC title

• G01H9/004 Primary

  using fibre optic sensors (light guides per se G02B6/00, acousto-optical devices specially adapted for gating or modulating in optical wave guides G02F1/125) · CPC title