AMUSE: Advanced Methodologies to Understand Dairy Powder Stability and Promote Exports

The overall aim of AMUSE is to ensure growth of exports of dairy powders, through a validated tool . AMUSE has been designed to implement developments and have real world impact throughout the project and beyond.

Dairy powder is a stronghold of the Danish industry (total exports 150000 tn, euro 450m >11% growth). To maintain growth, manufacturers have to provide to emerging markets affordable, high quality, stable powders under challenging supply chains i.e. high temperature & relative humidity. Currently there is no industrially relevant methodology to predict quality degradation, presenting a significant limitation for growth.

To ensure growth of exports AMUSE will provide a validated platform to predict shelf life under real supply chains. AMUSE will bring together ARLA and Lactosan (dairy manufacturers) with Videometer (instrumentation), UoLorraine and UoCopenhagen (research & technology development). AMUSE will build new understanding of quality degradation occurring across different length scales for powders in real supply chains. Using this knowledge, we will build “Supply Chain Reactors” to monitor powder changes under controlled environmental conditions relevant to real supply chains using state of the art sensors. Data will feed into mathematical models (Virtual Powders) that combine mechanistic with data driven approaches to predict quality degradation in real supply chains.

Outcomes of AMUSE will be reactors, methods and mathematical models, that combined will provide robust shelf-life prediction to support ongoing and future dairy powder development. This will create jobs and 10-year value of euro 32m from West Africa & South East Asia and euro 75m from the wider market.

Research objectives

To decipher the interconnected phenomena occurring during storage and provide real world impact AMUSE will achieve the overall aim through the following objectives:

Virtual Powders: new modelling toolboxes to predict shelf life indicators for real world supply chains. These will include mechanistic models to describe key phenomena at multiple length scales combined with data driven, e.g. machine learning, to capture complexity that is not accounted. The models will need a relatively large volume of experimental points at different storage conditions. Model output will be quality deterioration, e.g., surface changes, oxidation, protein degradation, crystalline lactose.
“Supply Chain Reactor” (SCR): experimental toolboxes that will provide the large volume of data required for the virtual powders; identify and use accelerated testing conditions. SCR will test powders under well controlled environmental conditions, i.e., modulations of temperature & relative humidity. SCR will be connected with relevant in-line/on-line sensors, e.g. NIR, LF-NMR, hyperspectral and will be eventually automated & robotised.
Real World Integration & Implementation: Characterise real world samples to support on-going product development. Integrate developments, e.g. models, stability protocols, in-line sensors in the current innovation pipeline. Improved logistics through reduced packaging; avoid cold storage; dynamic real shelf-life estimations by combinations with of weather forecast data.