Method and system to monitor and control continuous ultrafiltration process units

Patent metadata

Field	Value
Publication number	US-12472469-B2
Application number	US-202218049338-A
Country	US
Kind code	B2
Filing date	Oct 25, 2022
Priority date	Nov 2, 2021
Publication date	Nov 18, 2025
Grant date	Nov 18, 2025

Abstract

This disclosure relates generally to method and system to monitor and control continuous ultrafiltration (UF) process units. In real time, continuous operation of UF to handle variating concentration in feed stream is tedious and complex. The UF plant system receives a plurality of input data configured to UF process units and from the real time data outliers are removed and missing values are imputed. The prediction module predicts a volumetric concentration factor (VCF) value and a throughput value by selecting a model from a model repository. The optimization module optimizes the VCF value, and the throughput value based on a plurality of optimal variables recommended for a given feed concentration. The UF plant system controls the VCF value and the throughput value for a predefined period of a prediction horizon based on a plurality of trajectory profiles recommended for the feed flow rate, the pressure data, and a feed concentration.

First claim

What is claimed is:

1 . A processor-implemented method for monitoring and controlling continuous ultrafiltration (UF) process units, the method further comprising:
receiving, via a one or more hardware processors, a plurality of input data from one or more sensors configured to an UF process units, wherein the plurality of input data comprises a real time data and a non-real time data; pre-processing via the one or more hardware processors, the real time data by removing outliers and imputing missing values; converting, via the one or more hardware processors, an inline conductivity sensor data associated with the real time data into a concentration of protein of interest at (i) a feed flow stream, and (ii) a retentate stream of the UF process units based on a plurality of CDC models; predicting, via the one or more hardware processors, critical quality parameters (CQPs) comprises a volumetric concentration factor (VCF) value and a throughput value of the UF process units by selecting a model from a model repository using the real time data and the non-real time data further comprising (i) a pressure data, (ii) a feed flow rate and (iii) the concentration of protein of interest in the feed flow stream; optimizing, via the one or more hardware processors, the VCF value and the throughput value based on a plurality of optimal variables recommended for a given feed concentration, wherein the plurality of optimal variables comprises an optimal feed flow rate and an optimal pressure data, wherein a plurality of operating parameters of the UF process units is optimized, using a plurality of models from the model repository, to maximize or minimize or maintain the CQPs or a key performance indicators (KPIs) at a target value, wherein the CQPs or the KPIs of the UF process units comprises the VCF, the throughput, a fouling index and a time of operation of the UF process units, wherein the fouling index is an indicator of a remaining useful life (RUL) of a membrane, wherein the fouling index ranges from 0 to 10, where 0 represents no fouling as in case of fresh membrane, and 10 represents severe fouling such that the membrane is unable to concentrate a feed; controlling, via the one or more hardware processors, the VCF value and the throughput value for a predefined period of a prediction horizon based on a plurality of trajectory profiles recommended for the feed flow rate, the pressure data, and the feed concentration, retuning, via the one or more hardware processors, an optimization module using a self-optimization module, for a change observed on at least one of (i) constraint values, (ii) tolerance or convergence criteria of an optimization algorithm; selecting, via the one or more hardware processors, an optimal optimization algorithm by performing (1) changing an objective function, (2) changing the constraints values, (3) changing parameters the tolerance or the convergence criteria of the optimization algorithm, and (4) choosing an another optimization algorithm; recommending, via the one or more hardware processors, a trajectory of optimum values of a plurality of operating variables for a time period of a control horizon to the UF process units; performing, via the one or more hardware processors, a real-time dynamic optimization for a subsequent control horizon while a UF controller in the UF process units is implementing actuation profiles, and further controls, turns off pumps in the UF process units, when a pressure at the pump is more than a critical membrane pressure limit; and determining, via the one or more hardware processors, time stamps at which the pumps in the UF process units are turned on or turned off based on recommended optimal feed flowrate.

2 . The processor implemented method as claimed in claim 1 , wherein the real time data includes a transmembrane pressure data, the inline conductivity sensor data, the feed flow rate, and a tank level data.

3 . The processor implemented method as claimed in claim 1 , wherein the non-real time data includes an experimental measured value of concentration of protein of interest for the feed stream and the retentate stream.

4 . The processor implemented method as claimed in claim 1 , wherein the fouling index is considered as a constraint for the optimization module and a control module.

5 . The processor implemented method as claimed in claim 1 , wherein the VCF value and the throughput are controlled by,
estimate, the plurality of trajectory profiles for the VCF value and the throughput value with corresponding trajectories of the feed flow rate and the pressure data; and recommend, the plurality of trajectory profiles for the feed flow rate and the pressure data.

6 . The processor implemented method as claimed in claim 1 , wherein a prediction module is retuned by using a self-learning module, when a measured deviation between an experimental value of the VCF value and the determined VCF value exceeds a threshold of deviation.

7 . The processor implemented method as claimed in claim 1 , the one or more hardware processors are further configured by the instructions to:
detect, one or more faults in the UF process units based on the one or more expected profiles of VCF value by monitoring the VCF value associated with the plurality of operating variables for the change observed from a pre-defined range of values; identify, the root cause analysis on the one or more detected faults in the UF process units; recommend, one or more corrective actions based on a fault detection using historical data, wherein the historical data comprises information on a corrective action taken for specific faults in past instances; and recommend, the one or more corrective actions for the one or more faults detected.

8 . A system for monitoring and controlling continuous ultrafiltration (UF) process units, further comprising:
a memory storing instructions; one or more communication interfaces; and one or more hardware processors coupled to the memory via the one or more communication interfaces-, wherein the one or more hardware processors are configured by the instructions to:
receive, a plurality of input data from one or more sensors configured to an UF process units, wherein the plurality of input data comprises a real time data and a non-real time data;
pre-process, the real time data by removing outliers and imputing missing values;
convert, an inline conductivity sensor data associated with the real time data into a concentration of protein of interest at (i) a feed flow stream, and (ii) a retentate stream of the UF process units based on a plurality of CDC models;
predict, critical quality parameters (CQPs) comprises a volumetric concentration factor (VCF) value and a throughput value of the UF process units by selecting a model from a model repository using the real time data and the non-real time data further comprising (i) a pressure data, (ii) a feed flow rate and (iii) the concentration of protein of interest in the feed flow stream;
optimize, the VCF value and the throughput value based on a plurality of optimal variables recommended for a given feed concentration, wherein the plurality of optimal variables comprises an optimal feed flow rate and an optimal pressure data, wherein a plurality of operating parameters of the UF process units is optimized, using a plurality of models from the model repository, to maximize or minimize or maintain the CQPs or a key performance indicators (KPIs) at a target value, wherein the CQPs or the KPIs of the UF process units comprises the VCF, the throughput, a fouling index and a time of operation of the UF process units,
wherein the fouling index is an indicator of a remaining useful life (RUL) of a membrane,
wherein the fouling

Assignees

• Tata Consultancy Services Ltd

Inventors

• Buddhiraju Venkata Sudheendra
• Runkana Venkataramana
• Masampally Vishnu Swaroopji
• Premraj Karundev
• Kumar Vivek
• Sahu Swati

Classifications

• B01D61/145 Primary

  Ultrafiltration · CPC title

• B01D2313/701

  comprising a software program or a logic diagram · CPC title

• B01D2313/70

  Control means using a programmable logic controller [PLC] or a computer · CPC title

• B01D2311/24

  Quality control · CPC title

• B01D61/22 Primary

  Controlling or regulating · CPC title