Device and method for optical analysis using multiple integrated computational elements

Patent metadata

Field	Value
Publication number	US-2020018162-A1
Application number	US-201615540255-A
Country	US
Kind code	A1
Filing date	Sep 29, 2016
Priority date	Sep 29, 2016
Publication date	Jan 16, 2020
Grant date	—

Abstract

A method including generating integrated computational element (ICE) models and determining a sensor response as the projection of a convolved spectrum associated with a sample library with a plurality of transmission profiles determined from the ICE models. The method includes determining a regression vector based on a multilinear regression that targets a sample characteristic with the sensor response and the sample library and determine a plurality of regression coefficients in a linear combination of ICE transmission vectors that results in the regression vector. The method further includes determining a difference between the regression vector and an optimal regression vector. The method may also include modifying the ICE models when the difference is greater than a tolerance, and fabricating ICEs based on the ICE models when the difference is within the tolerance. A device and a system for optical analysis including multiple ICEs fabricated as above, are also provided.

First claim

The disclosure claimed is:

1 . A method, comprising:
generating a plurality of integrated computational element (ICE) models; determining a sensor response from a projection of a plurality of ICE transmission vectors associated with the plurality of ICE models and a convolved spectrum associated with a sample library; determining a regression vector based on a multilinear regression that targets a sample characteristic from the sample library and the sensor response; determining a plurality of regression coefficients in a linear combination of the plurality of ICE transmission vectors that results in the regression vector; determining a difference between the regression vector and an optimal regression vector associated with the sample characteristic; modifying the plurality of ICE models when the difference between the regression vector and the optimal regression vector is greater than a selected tolerance; and fabricating a plurality of ICEs based on the plurality of ICE models when the difference between the regression vector and the optimal regression vector is within the selected tolerance.

2 . The method of claim 1 , wherein generating the plurality of ICE models comprises selecting a random number of layers and a random thickness for each layer in each ICE model.

3 . The method of claim 1 , wherein modifying the plurality of ICE models comprises:
modifying a number of layers and a thickness for each layer for at least one of the plurality of ICE models to obtain a plurality of modified ICE models; determining a modified sensor response from the plurality of modified ICE models and the convolved spectrum associated with the sample library; determining a modified regression vector from the plurality of modified ICE models and the modified sensor response; determining a difference between the modified regression vector and the optimal regression vector; iteratively repeat the modifying the number of layers, the determining a modified sensor response, the determining a modified regression vector, and the determining a difference between the modified regression vector and the optimal regression vector until the difference between the modified regression vector and the optimal regression vector is within the selected tolerance; and fabricating the plurality of ICEs based on the plurality of modified ICE models.

4 . The method of claim 1 , wherein determining a difference between the regression vector and the optimal regression vector comprises determining a mean square error between the regression vector and the optimal regression vector.

5 . The method of claim 1 , further comprising determining the optimal regression vector from a partial least squares model of the convolved spectrum and the sample library, targeting the sample characteristic.

6 . The method of claim 1 , wherein determining the regression vector and the plurality of regression coefficients comprises calibrating an optical computing device with the plurality of ICEs and with the sample library.

7 . The method of claim 1 , further comprising determining an accuracy and a sensitivity of an optical computing device that includes the plurality of ICEs based on the regression vector and the sample library.

8 . The method of claim 1 , further comprising storing in a memory the plurality of regression coefficients for each ICE transmission vector when the difference between the regression vector and the optimal regression vector is within the selected tolerance.

9 . The method of claim 1 , further comprising:
measuring a spectral performance of each of the ICEs in a fabrication batch of the plurality of ICEs; selecting a combination of ICEs from the fabrication batch based on the spectral performance; and disposing the combination of ICEs in an optical computing device that measures the sample characteristic.

10 . The method of claim 1 , wherein fabricating the plurality of ICEs comprises measuring a performance of a post-fabrication combinatorial configuration between different ICEs from a fabrication batch for each of the plurality of ICEs.

11 . The method of claim 1 , wherein fabricating the plurality of ICEs comprises:
fabricating one or more of the plurality of ICEs sequentially; and re-modeling an ICE that has not been fabricated based on a post-fabrication spectral performance of the one or more of the plurality of ICEs.

12 . A device, comprising:
at least two integrated computational elements (ICEs) positioned to optically interact with sample light to generate a first modified light from a first ICE and a second modified light from a second ICE; and a detector that separately measures the first modified light to provide a first signal and the second modified light to provide a second signal, wherein each one of the at least two ICEs comprises a plurality of alternating layers of material and each layer of material has a thickness selected such that a linear combination of the first signal with the second signal is proportional to a sample characteristic.

13 . The device of claim 12 , further comprising a multiplexer that directs a first portion of sample light to the first ICE and a second portion of sample light to the second ICE.

14 . The device of claim 12 , wherein the detector comprises a first detector to measure the first modified light, and a second detector to measure the second modified light, the first detector being spatially separated from the second detector.

15 . The device of claim 12 , wherein the detector measures the first modified light and the second modified light separated in time.

16 . A system, comprising:
a light source that generates an illumination light to interact with a sample and form a sample light; an optical computing device comprising:
at least two integrated computational elements (ICES) positioned to optically interact with the sample light to generate a first modified light from a first ICE and a second modified light from a second ICE; and
a detector that separately measures the first modified light to provide a first signal and the second modified light to provide a second signal, wherein each one of the at least two ICEs comprises a plurality of alternating layers of material and each layer of material has a thickness selected such that a linear combination of the first signal with the second signal is proportional to a sample characteristic; and
a controller comprising a processor and a memory, wherein the processor forms the linear combination of the first signal and the second signal based on at least two regression coefficients associated with the at least two integrated computational elements (ICEs) stored in the memory.

17 . The system of claim 16 , wherein the optical computing device further comprises a multiplexer that directs a first portion of the sample light to the first ICE and a second portion of the sample light to the second ICE.

18 . The system of claim 16 , wherein the detector in the optical computing device comprises a first detector to measure the first modified light, and a second detector to measure the second modified light, the first detector being spatially separated from the second detector.

19 . The system of claim 16 , wherein the detector in the optical computing device measures the first modified light and the second modified light separated in time.

20 . The system of claim 16 , wherein the at least two ICEs comprise more than two ICEs but less than a number of principal components in a partial least squar

Assignees

• Halliburton Energy Services Inc

Inventors

• Price James M
• Dai Bin
• Jones Christopher Michael

Classifications

• G01J3/0227

  using notch filters · CPC title

• G01J3/027

  Control of working procedures of a spectrometer; Failure detection; Bandwidth calculation · CPC title

• G01N2201/129

  Using chemometrical methods · CPC title

• G01N2021/855

  Underground probe, e.g. with provision of a penetration tool · CPC title

• G01N21/84

  Systems specially adapted for particular applications · CPC title