Soft Sensor for Estimation and Identification of Reduced Dimensional Quality Control Inputs

Abstract

Abstract
Advances in instrumentation technology have equipped us with better process controlling set-ups for error detection and control that occurs in the industrial process plants. This in turn generates a large amount of data that is not always information rich. Additional sensor like a soft sensor can be used to modify the sensor to generate information rich data. Soft sensor are computational models that aid in the continuous or partial estimation of some process parameters which otherwise are difficult to measure using hardware sensors. In this paper, a soft sensor is modelled for a waste water treatment plant where principal component analysis is used to reduce the high dimensional data that is available from the hardware sensor to a low dimensional information rich data and also to identify the type of measured data. This can be used in conjunction with the sensor in the plant to improve the control performance and yield better product quality.
Keywords: WWTP, soft sensors, principal component analysis, latent vectors, scores

Cite this Article

Barasha Mali, S. H. Laskar. Soft Sensor for Estimation and Identification of Reduced Dimensional Quality Control Inputs. Journal of Instrumentation Technology & Innovations. 2017; 7(3): 24–29p.