Evaluation of malted barley with different degrees of fermentability using the Rapid Visco Analyser (RVA)

Visser, Magdalena Johanna (2011-03)

Thesis (MSc Food Sc)--University of Stellenbosch, 2011.

Thesis

ENGLISH ABSTRACT: The relationship between malt fermentability and rheological variables, measured by means of the Rapid Visco Analyser (RVA) and application of multivariate data analysis, was investigated. The RVA Kilned Malt method was optimised to achieve maximum rheological discrimination between malt samples, differing in fermentability. Five concentrations and two particle sizes were used to investigate each malt sample. Data were analysed by two different data analysis techniques, namely principal component analysis (PCA) and analysis of variance (ANOVA). Rheological variables for peak-height, -width, -area and time occurrence, were able to discriminate between high (Metcalfe, Flagship), intermediate (SSG 585, PUMA) and low malt fermentability (SSG 506, SSG 564). Variation in particle size showed insignificant (P>0.05) fermentability discrimination. The malt to water ratio of 1:1.5 provided the best discrimination in fermentability. PCA applied to the entire dataset was the superior data analysis technique. Partial least squares (PLS) regression and Soft Independent Modeling of Class Analogy (SIMCA) were applied to predict malt fermentability. Recorded RVA data was regressed with both apparent attenuation limit (AAL) and free amino nitrogen (FAN), independently. Developed PLS calibration models were validated by test set and segmented crossvalidation for AAL and FAN, respectively. The SIMCA classification model developed was based on different malt fermentability classes, each PCA validated independently by test set validation. A strong correlation between RVA analysis and AAL was obtained (r=0.92), while FAN delivered a weak correlation (r=0.59). Regarding the SIMCA model; the proportion of test set samples correctly classified in terms of malt fermentability was 83%. South African malt blends were predicted to have low malt fermentability. Simulated blends were predicted to have high fermentability when using a minimum of 80% Metcalfe blended with SSG 506. Blends containing higher percentages of the low malt fermentability cultivar (SSG 506) were predicted to have an overall intermediate fermentability. Different experimental conditions were investigated during RVA analysis (i.e. instrument model; time/temperature profile, enzyme activity and heating/cooling rate). Rheological measurement using different RVA models gave similar PCA results, indicating adequate sensitivity of the older instrument for discrimination purposes. Matching the time/temperature profile used in the commercial brewery mashing process was rejected due to increased analysis time and rheological noise while reducing fermentability discrimination. Inactivating malt enzymes prior to RVA analysis provided useful sample information, such as the large starch granule’s mean diameter, extract and starch content, by measuring peak height and iv time to peak. The amount of starch damage inflicted on a malt sample increased after repeated centrifugal milling, but was unaffected by the sieve size used. Multivariate data analysis is a suitable statistical technique applied to rheological data and provided more relevant information than traditional univariate techniques. The RVA can be considered an ideal instrument within a grain laboratory as it allowed the investigation of different operating conditions. It is beneficial to use an inexpensive, routine method of analysis to measure various interacting factors. RVA rheological measurement demonstrated to be a decisive monitor of malt fermentability and is highly recommended to be incorporated within the barley breeding, malting and brewing industries.

AFRIKAANSE OPSOMMING: Die verwantskap tussen mout fermenteerbaarheid en reologie-veranderlikes, gemeet met die “Rapid Visco Analyser” (RVA) en toepassing van meerveranderlike data analise is ondersoek. Die RVA “Kilned Malt” metode is geoptimeer om maksimum reologiese diskriminasie, tussen gars kultivars van verskillende fermenteerbaarheid, te lewer. Vyf konsentrasies en twee partikel groottes is gebruik in die ondersoek vir elke mout monster. Data is deur beide hoof komponent analise (HKA) en variansie-analise (ANOVA) ondersoek om die verskillende data analise metodes met mekaar te vergelyk. Reologiese veranderlikes vir piek-hoogte, - wydte, -area en -vormingstyd, kon diskrimineer tussen hoë (Metcalfe, Flagship), intermediêre (SSG 585, PUMA) en lae (SSG 506, SSG 564) mout fermenteerbaarheid. Variasie in partikel grootte kon nie beduidende diskriminasie in fermenteerbaarheid aantoon nie. Die mout-totwater konsentrasie van 1:1.5 het die beste diskriminasie in fermenteerbaarheid gelewer. Die toepassing van HKA op die hele datastel was die beter analitiese tegniek. Parsiële kleinste kwadrate (PKK) regressie en Sagte Onafhanklike Modellering van Klas Analogie (SIMCA) is toegepas om mout fermenteerbaarheid te voorspel. Regressie tussen RVA data en skynbare attenuasie limiet (AAL), sowel as vrye amino stikstof (FAN) inhoud, is afsonderlik uitgevoer. Die geldigheid van regressie modelle is deur middel van toets stel en gesegmenteerde kruis-validasie vir AAL en FAN onderskeidelik uitgevoer. SIMCA klassifikasie modelle is gebaseer op verskillende mout fermenteerbaarheids-klasse, waarvan elke HKA klas individueel geldig is. RVA analise het ‘n sterk korrelasie met AAL (r=0.92), maar ‘n swak korrelasie met FAN (r=0.59) getoon. Die SIMCA model het 83% van toets stel monsters as korrek geklassifiseer in terme van mout fermenteerbaarheid. Suid Afrikaanse mout mengsels is voorspel as swak fermenteerbaar. Nagebootste mengsels is voorspel as hoogs fermenteerbaar wanneer minimum 80% Metcalfe met SSG 506 vermeng word. Sodra ‘n hoër persentasie van die swakker fermenteerbaarheids-kultivar (SSG 506) bygevoeg is, word intermediêre fermenteerbaarheid voorspel. Tydens RVA analise is verskillende eksperimentele toestande ondersoek (byvoorbeeld instrument model; tyd/temperatuur profiel; ensiem aktiwiteit en verhittings/verkoelings tempo). Die gebruik van verskillende RVA modelle het soortgelyke HKA resultate gelewer. Dus bevat die ouer instrument aanvaarbare sensitiwiteit vir diskriminasie doeleindes. Nabootsing van die tyd/temperatuur profiel in die kommersiële brouproses is verwerp, aangesien analise tyd en reologiese geraas toegeneem het, terwyl fermenteerbaarheidsdiskriminasie verminder het. Inaktivering van mout ensieme voor RVA analise lewer nuttige monster inligting; deur veranderlikes soos piek-hoogte en piek-tyd te meet, kan die groot stysel korrel se gemiddelde deursnit, ekstrakwaarde en stysel inhoud verkry word. Herhaalde sentrifugale maling van ‘n mout monster lei tot beskadiging van stysel, maar dit word nie deur sif grootte beïnvloed nie. Die toepassing van meerveranderlike data analise op reologiese data is waardevol en lewer meer relevante inligting in vergelyking met tradisionele eenveranderlike data analise. Die RVA is ‘n ideale instrument vir gebruik in ‘n graan laboratorium aangesien dit verskillende operatiewe kondisies kan ondersoek. Die gebruik van ‘n enkele, goedkoop, roetine analitiese metode is voordelig en het die potensiaal om ‘n magdom interaktiewe faktore te meet. RVA reologiese meting demonstreer die vermoë as ‘n deurslaggewende tegniek vir die bepaling van mout fermenteerbaarheid, gevolglik word toepassing sterk aanbeveel binne die gars teëlings-, vermoutings- en brouers-industrieë.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/6650
This item appears in the following collections: