Extracting features for eis monitoring systems

Abstract

Systems and methods are provided for operating an electrolyzer. The systems and methods perform operations comprising extracting, by monitoring circuitry coupled to a plurality of electrolytic cells of an electrolyzer, a set of features comprising at least one of a goodness-of-fit measurement or direct current (DC) measurement of the plurality of electrolytic cells; tracking changes to the set of features over a time period; and generating, based on the changes to the set of features, a model representing operating conditions of the electrolytic cells on an individual electrolytic cell basis.

What is claimed is: 
     
         1 . A system that includes an electrolyzer comprising a plurality of electrolytic cells, each of the electrolytic cells comprising an electrolyte, two electrodes and a pair of bipolar plates, the system comprising:
 monitoring circuitry coupled to the plurality of electrolytic cells, the monitoring circuitry configured to perform operations comprising:
 extracting a set of features comprising at least one of a goodness-of-fit measurement or direct current (DC) measurement of the plurality of electrolytic cells; 
 tracking changes to the set of features over a time period; and 
 generating, based on the changes to the set of features, a model representing operating conditions of the electrolytic cells on an individual electrolytic cell basis. 
   
     
     
         2 . The system of  claim 1 , wherein the model is configured to estimate at least one of state-of-health or performance of a given electrolytic cell, predict whether the given electrolytic cell is operating and degrading normally or abnormally, or identify an abnormality of the electrolytic cell. 
     
     
         3 . The system of  claim 1 , wherein the model comprises a machine learning technique that is trained based on training data to predict health of an electrolytic cell, the training data comprising a plurality of training samples of Electrochemical Impedance Spectroscopy (EIS) data and associated performance or failure information. 
     
     
         4 . The system of  claim 1 , wherein the monitoring circuitry comprises an Electrochemical Impedance Spectroscopy (EIS) measurement system, the EIS generating an impedance as a function of frequency of each of the plurality of electrolytic cells. 
     
     
         5 . The system of  claim 4 , wherein the EIS generates the impedance over a range of frequencies from 0.1 mHz to 10 kHz, a subset of frequencies within the range of frequencies, or one or more specific frequencies within the range of frequencies. 
     
     
         6 . The system of  claim 1 , wherein the operations further comprise:
 processing an output of the model by a classifier to generate a label indicating whether operation of the electrolyzer corresponds to a baseline cell operation or an anomalous cell operation.   
     
     
         7 . The system of  claim 6 , wherein the classifier comprises a feature statistics process that uses a baseline training data set to fit each of the extracted features with a Gaussian model. 
     
     
         8 . The system of  claim 7 , wherein the operations comprise:
 determining that an individual feature of the extracted features is more than a threshold distance away from a center of the Gaussian model; and   in response to determining that the individual feature of the extracted features is more than the threshold distance away from the center of the Gaussian model, determining that the electrolyzer corresponds to the anomalous cell operation.   
     
     
         9 . The system of  claim 6 , wherein the classifier comprises a Random Forest Classifier comprising a decision tree that is used to generate the label. 
     
     
         10 . The system of  claim 6 , wherein the classifier comprises a Gaussian Mixtures classifier. 
     
     
         11 . The system of  claim 1 , wherein the goodness-of-fit measurement represents how well the electrolyzer fits an equivalent circuit model (ECM). 
     
     
         12 . The system of  claim 1 , wherein the DC measurement comprises a DC current or DC voltage. 
     
     
         13 . A method comprising:
 extracting, by monitoring circuitry coupled to a plurality of electrolytic cells of an electrolyzer, a set of features comprising at least one of a goodness-of-fit measurement or direct current (DC) measurement of the plurality of electrolytic cells;   tracking changes to the set of features over a time period; and   generating, based on the changes to the set of features, a model representing operating conditions of the electrolytic cells on an individual electrolytic cell basis.   
     
     
         14 . The method of  claim 13 , wherein the model is configured to estimate at least one of state-of-health or performance of a given electrolytic cell, predict whether the given electrolytic cell is operating and degrading normally or abnormally, or identify an abnormality of the electrolytic cell. 
     
     
         15 . The method of  claim 13 , wherein the model comprises a machine learning technique that is trained based on training data to predict health of an electrolytic cell, the training data comprising a plurality of training samples of Electrochemical Impedance Spectroscopy (EIS) data and associated performance or failure information. 
     
     
         16 . The method of  claim 13 , wherein the monitoring circuitry comprises an Electrochemical Impedance Spectroscopy (EIS) measurement system, the EIS generating an impedance as a function of frequency of each of the plurality of electrolytic cells. 
     
     
         17 . The method of  claim 16 , wherein the EIS generates the impedance over a range of frequencies from 0.1 mHz to 10 kHz, a subset of frequencies within the range of frequencies, or one or more specific frequencies within the range of frequencies. 
     
     
         18 . The method of  claim 16 , further comprising:
 processing an output of the model by a classifier to generate a label indicating whether operation of the electrolyzer corresponds to a baseline cell operation or an anomalous cell operation.   
     
     
         19 . The method of  claim 18 , wherein the classifier comprises a feature statistics process that uses a baseline training data set to fit each of the extracted features with a Gaussian model. 
     
     
         20 . A non-transitory computer-readable medium comprising instructions that, when executed by one or more processors, configure the one or more processors to perform operations comprising:
 extracting, by monitoring circuitry coupled to a plurality of electrolytic cells of an electrolyzer, a set of features comprising at least one of a goodness-of-fit measurement or direct current (DC) measurement of the plurality of electrolytic cells;   tracking changes to the set of features over a time period; and   generating, based on the changes to the set of features, a model representing operating conditions of the electrolytic cells on an individual electrolytic cell basis.