DEVELOPMENT OF A MICRO-NIRS- APP-CLOUD SYSTEM FOR ON-SITE DETERMINATION OF THE NUTRITIONAL COMPOSITION OF POULTRY FEED AND ITS VARIOUS INGREDIENTS

Precision feeding can improve poultry nutrition and at the same time reduce the excretion of surplus nutrients, thereby improving the sustainability of poultry production while reducing feed costs. Near infrared spectroscopy (NIRS) is a useful tool for precision feeding as it allows to accurately analyze feed and feed ingredients and enables precise formulation of rations according to known nutritional requirements, eventually avoiding over-formulation. The purpose of the Pocket Feed Lab project is to develop a portable micro-NIRS for the on-site analysis of poultry feed and poultry feed ingredients. The micro-NIRS sensor developed by aikemy GmbH is based on MEMS (micro electro mechanical system) technology. It connects with a mobile app via Bluetooth and finally to cloud services for data analysis and storage. Results are displayed in near-real time as percentage values in the app. Furthermore, the comparison of the results with user pre-specified test targets provides an immediate pass/fail score for each nutrient. The potential of three different MEMS, covering different regions of the NIRS wavelength range, was explored for use in the final dual-MEMS design. Calibrations are being developed to predict crude protein, crude fiber, crude fat, water soluble carbohydrates, crude ash and moisture. Over 1000 samples (poultry feed and its various ingredients: corn, wheat, soybean meal/press cake and distillers’ grain) were collected in Switzerland, different European countries and the USA. They were analyzed according to official methods at the accredited analytical laboratory of Agroscope. The training sample set covers the protein content recommendations suggested by different research centers and is in line with the nutrient composition of available commercial products reported by the Swiss poultry sector. In general, the statistics of the preliminary prediction models show a good potential of the micro-NIRS system in terms of coefficient of determination (R2), root mean square error of prediction(RMSEP), and ratio of the standard deviation of the validation set to the standard error of prediction (RPD). The results suggest that this system will be a useful tool for precision feeding in poultry production. Based on the current knowledge, the application of this device could be extended to additional nutritional parameters as well as other feeds and feed ingredients.