Efficient Flow Control in Industrial Processes: Estimating In-flow at Control Valves Without Physical Sensors

Bhagya R. Navada, Santhosh Krishnan Venkata*, Preeti Mohanty

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

In industrial processes, achieving an efficient flow is paramount as various loops are controlled through flow control. To optimize this flow, it is essential to monitor various influential components such as control valves, flow sensors, as well as input and output parameters to the valve. To maintain the desired flow rate at the outlet of the control valve, it's imperative that the valve's inflow remains higher than the desired output flow rate. In most applications, fluctuations in the inflow rate are typically not considered as a significant factor affecting changes in the output flow which is a key driving component leading to our research objective. This research paper's contribution is estimation of control valve's inlet flow rate without the need for physical sensors and validation of estimated flow rate. This estimator relies on an outflow measurement from an orifice flow meter and employs a first order process with dead time model to deduce the inflow to the control valve. The process model is formulated through data driven system identification, employing the input-output characteristics of the system. Furthermore, a pole-placement-based estimator is developed utilizing real-world data. The novelty of the approach is estimation of the flow rate at valve input which is overlooked by many researchers. To validate the performance of this estimator, it is deployed to compute the inflow in a real-life practical system. The results reveal a root mean square error of 0.029, signifying the accuracy and reliability of the designed estimator. The estimator performed better in terms of reduced root mean square error when compared to other methods.

Original languageEnglish
Pages (from-to)875-883
Number of pages9
JournalJournal of Robotics and Control (JRC)
Volume5
Issue number3
DOIs
Publication statusPublished - 2024

All Science Journal Classification (ASJC) codes

  • Control and Systems Engineering
  • Artificial Intelligence

Fingerprint

Dive into the research topics of 'Efficient Flow Control in Industrial Processes: Estimating In-flow at Control Valves Without Physical Sensors'. Together they form a unique fingerprint.

Cite this