Prediction of Mass Flow Rate in Pneumatic Conveying Using a System Identification Modeling Approach

Abstract

Accurate measurements of solids flow rate in pneumatic conveying systems have been a need for the different industries in which these systems are used. In this study, a system identification approach, which is called the Deterministic Stochastic Realization (DSR) method, is used to obtain a state-space model. The selection of inputs to the system identification model is based on conservation of energy related to the Bernoulli principle. Under this investigation, attempts were made to predict the mass flow rate of conveying solids in real time by using pressure data obtained from three different pressure sensors located on the conveying line, blow tank pressure, and total air volume flow rate. Tests were performed in two different pneumatic conveying rigs and with two different materials. The conveyed materials were baryte and dextrose. Dextrose and baryte were conveyed in dilute and dense phases, respectively. To evaluate the DSR model, the results were compared with one published method based on a scaling up technique of pneumatic conveying. The root mean squared error of prediction (RMSEP) of the cumulative mass of transported solids as a percentage of the total conveyed mass using the system identification model was between 1.4 and 5.6% for baryte and 3.2 and 7.8% for dextrose. The cumulative error was between 0.7 and 7.2% for baryte and 3.7 and 8.6% for dextrose.