Designing of Mass Flow Observer System of a Mineral Processing Plant

"The main objective of this study is to find a solution to design a mass flow observer system with the minimum required sensors and maximum observability for a mineral processing plant. The second objective is designing a mass flow observer configuration with high reliability to manage sensor failures. This configuration will provide a robust monitoring system that increases plant availability and process measurement fault detection. Furthermore, the monitoring system is expected to improve process efficiency. In this investigation, the procedure of designing an observer system is presented along with three models to design the sensor network. The advantages and disadvantages of each models are discussed and finally the models result for an industrial case are presented and analyzed. INTRODUCTION The concepts of observability and redundancy of mass flow are directly linked with the ability to solve and estimate the mass flow. There are hundreds of streams in a process plant flowsheet, and for technical and economic reasons it is not possible to measure all mass flows. It is thus important to know for a set of measured streams, which of the unmeasured streams can be estimated. The concept of observability deals with this issue. It is also valuable to know whether a measured mass flow of streams can be estimated even if a sensor fails for some reason. Redundancy deals with this failure question. A stream is said to be observable if it can be estimated by using the measurements and/ or mass conservation equations. Also, a stream is redundant if it is observable even when its measurement is removed. Observability and redundancy analysis can be exploited for adding new measuring instruments. It can also play a useful role in efficient decomposition and data reconciliation problem (Narasimhan and Jordache 1999). The value of an observer system cannot be justified without considering its benefits. This justification is done through evaluation of data reconciliation, gross error detection and improving process performance such as outlined below:"