Characterising the Rheology of Fermented Dairy Products During Filling

Abstract (Swedish)

Fermenterade mejeriprodukter har utvecklats i flera samhällen, och idag är yoghurt den mest populära fermenterade mejeriprodukten över hela världen. En av svårigheterna med bearbetning och modellering av fermenterade mejeriprodukter är deras komplexa beteende, som orsakas av den gelstruktur som bildas under fermentering samt deras anisotropiska natur.

För Tetra Pak är de enklaste reologimodellerna inte tillräckliga som beteendeindikatorer eller för modellering av fermenterade mejeriprodukter. Därför är det nödvändigt att undersöka vilka modeller och mättekniker som bättre kan beskriva fermenterade mejeriprodukter. Den här studien ämnade att förbättra karakterisering av fermenterade mejeriprodukter genom att tillämpa en mängd existerande mättekniker. Tre produkter karakteriserades med de studerade mätteknikerna. Resultaten korrelerades till problem som identifierades under tester utförda i en fyllrigg. Reologimodellerna validerades genom att mäta produkternas yt- och tjockleksprofil på en metallplatta med en laserskanner (‘metallplatta-riggen’), och jämfördes med CFD-simuleringar.

Den här studien demonstrerade att det är möjligt att mäta en märkbar skillnad mellan produkterna med valda mättekniker och analysmetoder, samt existerande utrustning. Testerna i metallplatta-riggen demonstrerade att det är möjligt att mäta yt- och tjockleksprofilen av produkterna. Det var inte var möjligt att jämföra resultaten med CFD-simulationerna på grund av komplikationer som uppstod under modelleringen. Testerna i fyllriggen demonstrerade att det var möjligt att kvantifiera problem som kan uppstå i automatiska fyllmaskiner, såsom ‘droppande’ och ‘stänk’. De påföljande statistiska analyserna (Pearson korrelation och Partial Least Square) tyder på att frekvensen av dessa problem kan förutses genom användning av mätbara reologiska parametrar. Baserat på resultaten från karakteriseringen och testerna i fyllriggen kunde slutsatsen dras, att variablerna A, b och nε kan mätas och korreleras till de studerade produkternas flödesbeteende.

Följande mättekniker och analysmetoder föreslås för karakterisering av de studerade produkterna:
• Linjär stress ramp och Tangent-analysmetod för bestämning flytgräns
• Skjuvhastighetssvep och Herschel-Bulkley kurvanpassning (flytgräns inkluderad) för bestämning av A och b
• Metallplatta valideringsmetod för jämförelse med framtida CFD modellering
• Fyllrigg tester i pilotskala för identifiering och kvantifiering av problem som uppstår under automatisk fyllning

Popular Abstract

It is difficult to predict the flow behaviour of yoghurt during processing because it is a complex fluid that is problematic to characterise. This study correlated measurable flow properties to issues during automatic filling.

Yoghurt is currently the most popular fermented dairy product world wide. The complexity of yoghurt causes its viscosity, the property describing resistance to flow, to change with applied force and time. Currently, Tetra Pak uses a fluid model called ‘Power Law’, which is not sufficient for describing the complex behaviour of yoghurt. Because of this, Tetra Pak expressed an interest in investigating if the measurable flow properties of three different products could be correlated to common issues observed during automatic filling.

The aim of the study was to perform a multivariate statistical analysis using new measurable parameters and issues, and hopefully find correlations between them. This would reveal which parameters could possibly be used for anticipation of specific issues during automatic filling. If issues better can be anticipated and the filling process optimised to minimise these issues, this could lead to a decrease in cleaning time and an increase in plant efficiency and capacity.

The products chosen for studying were Skånemejerier Vaniljyoghurt, Skånemejerier Naturell Lättyoghurt and Arla Långfil. The first two products are yoghurts with slightly different compositions. The last product is a Scandinavian sour milk and contains a fermentation culture that results in exceptionally long polysaccharides, causing it to exhibit extensional properties such as very ‘stringy’ behaviour.

Several fluid models exist to describe liquid food products, and includes different parameters. The most common parameters are those that describe the products consistency and flow behaviour. The parameters included in this study describing the consistency in different ways are K, Kε and A. The parameters included in this study describing the flow behaviour in different ways are n, nε and b. The parameters K and n come from the Power Law model. Kε and nε also come from the Power Law model, but instead describe the product’s extensional properties. Similarly, the parameters A and b also describe the fluid behaviour as in Power Law but with the addition of a yield stress (σy), which is part of the ‘Herschel-Bulkley’ model. Yield stress describes a fluid’s resistance to flow when exposed to very low forces. This behaviour is exemplified by toothpaste. When squeezed onto a plane surface, toothpaste will not move unless exposed to some minimum external force (the yield stress).

In addition to these seven parameters, two other parameters were also included in the study, both acting as boundary conditions of the product’s viscosity. Zero-shear viscosity (η0) describes the maximum viscosity of the products when exposed to very low forces, and viscosity at infinite shear rate (η∞) describes the minimum viscosity of the products when exposed to extreme forces.

Filing issues were quantified according to their frequency for each product. The following were defined as common issues during automatic filling:
Droplets. Small, round deposits of product outside of the designated fill volume.
Streaks. Longer deposits of product located outside of the designated fill volume.
Drips. The delayed formation of a droplet from product remaining on the fill nozzle after the end of the fill cycle that either falls outside of the designated fill volume.
Filaments. A ‘string’ of product reaching from the fill nozzle into the package, formed following the closure of the filling nozzle. This is counted when the filament is still present after the end of the fill cycle.

Droplets and streaks were only observed in the yoghurts, mostly in Naturell Lättyoghurt and never in Långfil. Filaments were observed at the end of every fill cycle for both yoghurts, and never in Långfil. Drips were never observed in the yoghurts but frequently in Långfil.
Statistical analysis found there were correlations between some of the issues and parameters. b and nε were positively correlated to all issues except drips. Conversely, K, A and Kε were negatively correlated to all issues except drips. Finally, σy, n and η0 appeared to be uncorrelated.
The strong, positive correlation of b and nε to all filling issues except drips is reasonable since they both describe the products’ flow behaviours. This meant that products with high values for b and nε also had more occurrences of these issues per ten filling shots. Therefore, all issues except drips were considered to be related to the products’ flow indices. Drips were instead considered to be related to the products’ extensional behaviour, since this phenomenon only occurred in Långfil. With this in mind it would seem reasonable that Kε and nε are strongly correlated to drips.

Together, A and b showed stronger correlations than K and n, to the observed issues, indicating that the Herschel-Bulkley model is more suitable for anticipating filling issues than Power Law. While σy showed no correlation on its own, it is required to determine A and b.

Based on the results in this exploratory study, Herschel-Bulkley is recommended as a more suitable model than Power Law for anticipating issues during automatic filling. Its parameters can potentially be used for anticipating specific issues during automatic filling, and also be used for optimisation of the filling process. This will hopefully lead to an increase in yoghurt filling efficiency and production capacity, and thus contribute to an overall improvement to the process.

• author: Williamson, Hera and Ostréus, Rebecka
• supervisor: Andreas Håkansson
• organization: Food Technology and Nutrition (M.Sc.)
• course: KLTM05 20191
• year: 2019
• type: H2 - Master's Degree (Two Years)
• subject: Technology and Engineering
• keywords: yoghurt, metodutveckling, method development, viskositet, viscosity, flytgräns, yield stress, Herschel-Bulkley, livsmedelsteknik, food engineering
• language: English
• id: 8984434
• date added to LUP: 2019-07-29 14:40:05
• date last changed: 2019-07-29 14:40:05