Viscosity Measurements for Milk Powder Process Control

Abstract

The aim of this work is to understand the main concept behind the potential use of viscosity measurement for milk powder process control and how the understanding of the dynamic of viscosity measurement can benefit the process. To achieve this goal, this study used the fundamentals of rheology, system identification, and a variety of viscosity measurement technologies with potential application for the dairy industry. Finally, a measurement application for the dairy industry was proposed. In order to choose one specific technology and strategy for viscosity measurement in the milk powder process, a comprehensive literature review of viscometry for the dairy industry, milk powder process control and milk rheology, was carried out. The potential importance of viscosity for the milk powder process and how it can relate to many of the quality parameters of the final product is also presented in this work. In addition, the way in which process control can affect, or even optimize, desired milk powder characteristics for a competitive commercial product is discussed. The milk powder process will be briefly described to situate the reader in the chosen application scenario for the viscometry problem of concern. In this study, the rheological behaviour of different milk materials was evaluated. The data were obtained through “off-line” tests, which were carried out using the Rheometer ARG2 that was available from the Chemical & Materials Department at The University of Auckland. An experimental rig was designed for this project; details of instruments mounted in the rig as well as the calculations performed are presented in this study. The requirements for the identification of potential use of viscosity for monitoring and process control are discussed in this work. Two experimental Programmes were designed for the project. In this thesis, an introduction is given to the materials, trial processes, details of the experiments, measurements taken, and the process variables and parameters for each of the Programmes. Finally, the methodology applied for system identification is also described. Initially, tests were carried out at the University of Auckland using reconstituted milk. Later on, milk from Fonterra's Te Rapa site was analysed. The rig was transported to Waikato and set up at the calibration lab in the Advance Process Control building at Te Rapa. Furthermore, in this report, an evaluation of the suitability of the different single input and single output (SISO) and multiple input and single output (MISO) models identified for the data gathered are presented. Autoregressive, state-space and transfer functions models were used to identify and evaluate the dynamic response of viscosity for different experimental conditions. Finally, a methodology to measure viscosity using a Coriolis meter is introduced and the results are presented. This new viscosity measurement technique is based on work developed for the pulp and paper industry. Inputs for the controller design are introduced and discussed.