Prediction of flow by linear parameter varying model under disturbance

Flowrate being one of the most measured process parameters, thus it is essential to produce higher accuracy. The objective of the paper is to design an estimator using the Linear Parameter Varying (LPV) model which would estimate the flow rate using the data from the orifice meter for varying parameters like, the beta ratio of an orifice, and liquid density. For the development of an estimator, a process model is used, which is designed with the help of a system identification technique using a data-driven method. Data for system identification is obtained by the process model designed using Computational Fluid Dynamics (CFD). The output of the CFD model is compared with experimental results, there is a good agreement in results obtained with an average of 5.97% and 3.19% error in terms of differential pressure and discharge coefficient respectively. The estimated output from the LPV model is compared with that of results obtained from the neural network model and experimental setup, which are also in good agreement with an average error of 2.14%. Thus can be used to estimate flow, when orifice design or density of fluid change intentionally or unintentionally.