A critical process analysis of wine production to improve cost efficiency, wine quality and environmental performance

Sheridan, Craig (2003-03)

Thesis (MScAgric)--Stellenbosch University, 2003.

Thesis

ENGLISH ABSTRACT: Wine cellars are diverse in terms of equipment types and process configurations. Whilst other food production processes have, in many cases, been properly analysed and modelled, this process diversity has resulted in an absence of process analyses in the wine industry. Each wine cellar is unique and represents a fully integrated agro-business, starting with a raw material (grapes) and extending to marketing and selling of the final product (wine). This makes the wine industry unique in this context. This study is the first attempt to analyse winemaking procedures in the form of a process audit. The study was approached in the following manner: • A questionnaire was developed to assess cellar configurations and conditions. This questionnaire was submitted to a statistically significant number of cellars, and a statistically significant number of questionnaires were returned. • The data collected from the questionnaire were statistically analysed and associations between equipment or procedures and wine faults were identified. • Three cellars were studied in depth. These three cellars had their processes audited and their effluent characterised. Additional data were obtained from current sampling projects and these data were analysed to complement the data obtained from the questionnaire • A preliminary input/output model was developed. The major results of this study are: • It was found that certain faults that appear in wine might be associated with equipment and/or process faults. These associations are statistically significant and they show that cellar hygiene is of critical importance when assessing these wine faults. The most important of these faults are VA, microbial contamination of the wine, sluggish and stuck fermentations. A risk hierarchy was derived to indicate which events are associated with others most strongly. • It was found that few wineries measure water consumption and even fewer wineries measure the quantity of effluent produced. • Correlations have been developed to predict winery parameters in terms of tons of grapes pressed per annum. These parameters include water and electricity consumed, wine produced and the quantity of effluent produced. Effluent characteristics have also been correlated to the tons of grapes pressed per annum. These characteristics include chemical oxygen demand, sodium absorption ratio and total dissolved solids in solution. Chemical oxygen demand was identified as the most important contributing factor in winery effluent. It was shown that all variables rise with an increase in cellar size, but the rise is not linear. This implies that large cellars have greater quantities of effluent of lower quality than small cellars. Most cellars have effluent concentrations that require some form of effluent treatment. The characterisation of effluent shows that the most widely used disposal practice is irrigation, and that the effluent disposed in this manner does not meet legislative requirements. • A preliminary input/output model was developed in order to enable wineries that have not measured the relevant parameters to predict the abovementioned variations. The resolution of these predictions is low but the model serves to provide an initial estimate if there are no data available. The model will give industrial averages for any given cellar size. • An economic balance was performed using this preliminary model. It was shown that if cellars were to lower the consumption of utilities and to reduce the strength of their effluent (using cleaner practices and not dilution) the reduction of operating costs could be reduced by 14% for smaller cellars to 17% for larger cellars. This study has shown that it is possible to make wine in a more environmentally friendly manner, producing better quality wines, without incurring extra costs.

AFRIKAANSE OPSOMMING: Wynkelders is diverse eenhede ten opsigte van die tipe toerusting en prosesse wat gebruik word. Terwyl ander prosesse rondom voedselproduksie in baie gevalle reeds geanaliseer en gemodelleer is, het die diversiteit in die wynindustrie gelei tot 'n afwesigheid van prosesanalises. Elke wynkelder is uniek en verteenwoordig 'n ten voile geintegreerde agro-besigheid wat begin met die rou materiaal (druiwe) en lei tot die bemarking en verkope van die finale produk (wyn), 'n eienskap wat die wynindustrie uniek maak in hierdie verband. Hierdie studie is die eerste poging om die wynmaakproses in die vorm van ‘n proses-oudit te analiseer. Die studie is soos volg aangepak: • ‘n Vraelys is ontwikkel om kelderkonfigurasies en -toestande te ondersoek. Die vraelys is aan ‘n statisties betekenisvolle aantal kelders voorgele, en ‘n statisties betekenisvolle aantal vraelyste is terugontvang. • Die data wat uit die vraelyste ontvang is, is statisties ontleed en verwantskappe tussen toerusting of prosesse en wyngebreke is geidentifiseer. • Drie kelders is in diepte bestudeer. Hierdie drie kelders se prosesse is geoudit en die afvloeiwater is gekarakteriseer. Addisionele data is verkry van huidige projekte en hierdie data is ge-analiseer om die data van die vraelys aan te vul. • ‘n Voorlopige inset / uitset model is ontwikkel. Die belangrikste resultate van hierdie studie is: • Dit is bevind dat sekere gebreke wat in wyn voorkom geassosieer kan word met tekortkominge in toerusting en/of prosesse in die kelder. Hierdie assosiasies is statisties betekenisvol en toon dat kelderhigtene van kritiese belang is wanneer gebreke in wyn ondersoek word. Die mees belangrike gebreke wat voorkom is vlugtige suur, mikrobiese kontaminasie van wyn, slepende en gestaakte fermentasie. ‘n Risikohierargie is afgelei om die gebeure te toon wat die sterktste met mekaar geassosieer word. • Dit is gevind dat min wynkelders waterverbruik meet. Selfs minder kelders meet die hoeveelheid afvloeiwater wat geproduseer word. • Korrelasies is ontwikkel om kelderparameters te voorspel in terme van ton druiwe gepars per jaar. Hierdie parameters sluit in water- en elektrisiteitsverbruik, wyn geprosuseer en hoeveelheid afvloeiwater geproduseer. Eienskappe van afvloeiwater is ook gekorreleer met die ton druiwe wat per jaar gepars word. Hierdie eienskappe sluit in chemiese suurstofbehoeftes, natrium absorpsieverhoudings en totale opgeloste soliede materiaal in oplossing. Chemiese suurstofbehoeftes is geidentifiseer as die mees belangrike bydraende faktor tot afvloeiwater in kelders. Dit is getoon dat alle veranderlikes verhoog hoe groter die kelder, maar hierdie verhoging is nie lineer nie. Dit impliseer dat groter kelder meer afvloeiwater van ‘n laer kwaliteit produseer. Die meeste kelders produseer afvloeiwater van sodanige konsentrasies dat behandeling daarvan nodig is. Die ondersoek van hierdie afvloeiwater toon dat die mees algemene wyse van wegdoening van afvloeiwater besproeiing is, en dat afvloeiwater wat op hierdie manier weggedoen word, nie voldoen aan die wetgewing se vereistes nie. • 'n Voorlopige inset/uitset model is ontwikkel om kelders wat nie die toepaslike parameters gemeet het nie in staat te stel om hierdie parameters te voorspel. Die akuraatheid van hierdie voorspellings is nie hoog nie, maar die model verskaf ‘n aanvanklike skatting waar daar geen data beskikbaar is nie. Die model verskaf industriele gemiddeldes aan kelders van enige grootte. • ‘n Ekonomiese balans is uitgevoer deur van hierdie model gebruik te maak. Dit is getoon dat indien kelders die gebruik van water en elektrisiteit verminder en die konsentrasie van afvloeiwater verlaag (deur van skoner paktyke gebruik te maak, en nie verdunning nie) die bestuurskoste met 14% vir kleiner kelders tot 17% vir groter kelders verlaag kan word. Die studie het getoon dat dit moontlik is om wyn te maak op ‘n meer omgewingsvriendelike wyse, en sodoende beter kwaliteit wyn te produseer sonder addisionele kostes.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/53239
This item appears in the following collections: