• Dorthe Kristensen
4. term, Chemical Engineering, Master (Master Programme)
Local, small-scale wineries can today employ fast, easy-to-use, inexpensive techniques for measuring chemical constituents for process and product control to guard against failed fermentations as early as possible and to further optimal wine quality. NIR is a reliable, fast, accurate, and precise method for measuring a. o. sugars, acids, glycerol, and ethanol using newly developed, robust, hand-held short- and long-wave NIR equipment. Two general PhazirTM sensor systems (short- and long-wave NIR), were employed, based on chemometric multivariate calibration on three low-alcohol wines made from locally harvested berry and fruits in a local winery, Tromsø, Norway. For each wine a PLS-regression calibration model for selected components of wine-making interest was optimized; all PLS-regression models were validated with test-set validation exclusively. Satisfactory individual models were obtained for sucrose, glucose, fructose, glycerol, ethanol, citric, tartaric, malic, succinic, lactic, and acetic acid, and excellent models for total sugar and total acid, when using the short-wave NIR region (890 – 1690 nm); 1.st derivative pre-treatment was applied to all parameters except for ethanol where full MSC is used. There is a high likelihood of carrying over these results also to table wines and fermentation broths.

Variographic analysis (Theory of Sampling) evaluates process variability and uncertainties associated with sampling of waste water. This study analysed both incremental samples and 24-hour composite samples from the Bramming Waste Water Treatment Facility North including total phosphorus (TP), conductivity, and ammonium (NH4-N). For 48 hours, short-term increments were extracted at two selected sampling points to evaluate the total sampling error and effects from periodicity; these increments were analyzed individually and their variogram showed a cyclic variation of approx 420 m3, which corresponds to day-to-night load variations. The 24 hour composite samples were concluded to be reliable results for emission monitoring. Also, for a period of 39 days, standard composite samples (24-hour intervals) were collected from seven different locations in the standard sampling outlet facility and compared. Variographic analysis of TP and conductivity from these alternative outlet locations result in a marked cyclic variation of 7 days (TP) and 9 days (conductivity). This cycle equals week vs. weekend variations in composition. The indicated 9-day period remains unexplainable. The present study confirms that the standard sampling point F, directly after the weir, is representative, but an equally good alternative was found in sampling point B. Total Sampling Error (TSE) for TP and conductivity could be estimated as 12.3 % and 12.4 % respectively, while the Total Analytical Error (TAE) was insignificant. Variographic analysis of NH4-N in increments indicates a 12-lag cyclic variation, interpreted as a possible pseudo-variation due to systematic analysis delays.
Publication date14 Jun 2010
Number of pages212
Publishing institutionAalborg Universitet Esbjerg
ID: 32926869