Method and system for estimating corrosion inhibitor concentration using a multi-electrode array sensor

Patent metadata

Field	Value
Publication number	US-11699505-B2
Application number	US-201916699518-A
Country	US
Kind code	B2
Filing date	Nov 29, 2019
Priority date	Nov 29, 2018
Publication date	Jul 11, 2023
Grant date	Jul 11, 2023

Abstract

There is a demand for low-cost robust method to detect corrosion for estimating corrosion inhibitor (CI) concentration sensing. This disclosure herein relates to method and system for estimating corrosion inhibitor (CI) concentration using a multi-electrode array sensor. The method initially obtains a plurality of electrochemical signals using the multi-electrode array sensor from the corroding environment. Further, the plurality of electrochemical signals are analyzed to obtain a plurality of parameters. Further, the method analyses a plurality of features from the plurality of parameters for estimating the corrosion inhibitor (CI) concentration using a trained machine learning model. The method is capable of estimating the corrosion inhibitor concentration of any unknown liquid using the regression model and the classification model.

First claim

What is claimed is:

1. A processor implemented method ( 300 ) for estimating corrosion inhibitor (CI) concentration, wherein the method comprises:
positioning ( 302 ), via the one or more hardware processors, a multi-electrode array sensor in a corroding environment for estimating a corrosion inhibitor (CI) concentration associated with a liquid media from the corroding environment, wherein the liquid media from the corroding environment is circulated externally into glassware by a pump, wherein the glassware contains holes to insert and fix the multi-electrode array sensor into the liquid media for in-situ estimation of CI concentration of the corroding environment;
obtaining ( 304 ), via the one or more hardware processors, from the corroding environment, a plurality of electrochemical signals using the multi-electrode array sensor, wherein the plurality of electrochemical signals are obtained using the multi-electrode array sensor having a sensing working electrode made of mild steel, a reference electrode made of a graphite stick or Ag—AgCI type electrode, and a counter electrode made of Pt mesh or graphite stick bundle, wherein the plurality of electrochemical signals is obtained by applying +0.1 V DC voltage into the working electrode with respect to the reference electrode, and a DC current is measured at the counter electrode;
analyzing ( 306 ), via the one or more hardware processors, the plurality of electrochemical signals to obtain a plurality of parameters comprising an electrochemical impedance spectra and a DC current versus time;
identifying ( 308 ), via the one or more hardware processors, a plurality of features from the plurality of parameters comprising:
a charge transfer resistance feature obtained from the electrochemical impedance spectra parameter, and
an initial DC current and an intermediate DC current obtained from the DC current versus time, wherein the initial DC current is measured at a predetermined elapsed time of 10 seconds, and the intermediate DC current is measured as an average DC current between a predetermined time range of 2000 to 3000 seconds;

estimating ( 310 ), via the one or more hardware processors, the corrosion inhibitor (CI) concentration from the plurality of features using a trained machine learning model comprising,
a regression model determining a quantitative estimate of the corrosion inhibitor (CI) concentration, wherein the regression model is trained using a training dataset utilized as a ground truth to determine the quantitative estimate of the corrosion inhibitor (CI) concentration, wherein training the regression model for determining the quantitative estimate of the corrosion inhibitor (CI) concentration comprises:
determining a reference corrosion inhibitor (CI) concentration for the regression model based on the training dataset, wherein the reference corrosion inhibitor (CI) concentration for the regression model varies between 0-0.5 mM (in milli molar unit);
computing a cost function using a pre-defined initial weightage factor θ corresponding to the plurality of features, the training dataset, and a measure of error for fitting the cost function to the reference corrosion inhibitor (CI) concentration of the regression model;
minimizing the computed cost function corresponding to the plurality of features, based on a learning rate α, a weightage factors θj for the plurality of features and the total number of iterations performed for the cost function;
determining a final weightage factors corresponding to the plurality of features; and
determining the quantitative estimate of the corrosion inhibitor (CI) of the plurality of features using the final weightage factors; and

a classification model determining a qualitative estimate of the corrosion inhibitor (CI) concentration, wherein the classification model is trained using the training dataset utilized as the ground truth to determine the qualitative estimate of the corrosion inhibitor (CI) concentration, wherein training the classification model for determining the qualitative estimate of the corrosion inhibitor (CI) concentration comprises:
determining a reference corrosion inhibitor (CI) concentration for the classification model based on the training dataset, wherein the reference corrosion inhibitor concentration for the classification model varies between 0-6.0 mM (in milli molar unit);
computing a cost function using a logistic function, a pre-defined initial weightage factors θ corresponding to the plurality of features, the training dataset and a measure of error for fitting the cost function to the reference corrosion inhibitor (CI) concentration of the classification model;
minimizing the computed cost function corresponding to the plurality of features, based on a learning rate (α), a weightage factors θj for the plurality of features and the total number of iterations performed for the cost function;
determining a final weightage factors corresponding to the plurality of features; and
determining the qualitative estimate of the corrosion inhibitor (CI) of the plurality of features using the final weightage factors.

2. The method as claimed in claim 1 , wherein the corrosion inhibitor concentration for any unknown liquid is estimated using the corrosion inhibitor concentration range of the regression model.

3. The method as claimed in claim 1 , wherein the corrosion inhibitor concentration for any unknown liquid is estimated using the corrosion inhibitor concentration range of the classification model.

4. A system ( 102 ) for electrical load disaggregation, the system ( 102 ) comprising:
a memory ( 202 ) storing instructions;
one or more Input/Output (I/O) interfaces ( 206 );
and one or more hardware processors ( 204 ) coupled to the memory ( 202 ) via the one or more I/O interfaces ( 206 ), wherein the one or more hardware processors ( 204 ) are configured by the instructions to:
position a multi-electrode array sensor in a corroding environment for estimating a corrosion inhibitor (CI) concentration associated with a liquid media from the corroding environment, wherein the liquid media from the corroding environment is circulated externally into a glassware by a pump, wherein the glassware contains holes and fix to insert the multi-electrode array sensor into the liquid media for in-situ estimation of CI concentration of the corroding environment;
obtain from the corroding environment, a plurality of electrochemical signals using the multi-electrode array sensor, wherein the plurality of electrochemical signals are obtained using the multi-electrode array sensor having a sensing working electrode made of mild steel, a reference electrode made of a graphite stick or Ag—AgCI type electrode, and a counter electrode made of Pt mesh or graphite stick bundle, wherein the plurality of electrochemical signals is obtained by applying +0.1 V DC voltage into the working electrode with respect to the reference electrode, and a DC current is measured at the counter electrode;
analyze the plurality of electrochemical signals to obtain a plurality of parameters comprising an electrochemical impedance spectra and a DC current versus time;
identify a plurality of features from the plurality of parameters comprising:
a charge transfer resistance feature is obtained from the electrochemical impedance spectra parameter,
an initial DC current and an intermediate DC current are obtained from the DC current vs time, wherein the initial DC current is measured at a predetermined elapsed time of 10 seconds, and the intermediate DC current is measured as an average DC current between a predetermined time range of 2000 to 3000 seconds; and

estimate the corrosion inhibitor (CI) concentration from the plurality of features using a trained machine learning model comprising,
a regression model determining a quanti

Assignees

• Tata Consultancy Services Ltd

Inventors

• Kanamarlapudi Venkata Muralidhar
• Dingari Naga Neehar
• Deshpande Parijat
• Dutta Jayita
• Maiti Soumyadipta
• Rai Beena

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G01N17/02 Primary

  Electrochemical measuring systems for weathering, corrosion or corrosion-protection measurement · CPC title

• G06N20/00

  Machine learning · CPC title

• G16C20/70 Primary

  Machine learning, data mining or chemometrics · CPC title

• G16C20/30

  Prediction of properties of chemical compounds, compositions or mixtures · CPC title