BOX-WINE: the quality of wine under control

The wine production becomes smart

Very often one wonders whether a product is really of quality or not, especially with regard to food and drink as it is directly responsible for any problems that may create to our health. This happens even more when you are facing a bottle or a glass of wine; accentuated by the fact that we are constantly looking for the right quality-price ratio.  

But what are the elements that make a wine a quality product and above all, what are the technologies in the hands of the wineries that allow to monitor all the different parameters?

We at EDALAB ci siamo chiesti come i sistemi IoT nelle vostre mani possano essere declinati per cercare di controllare e monitorare tutti i parametri che contribuiscono alla produzione di un vino di qualità. Il nostro intento è stato quello di realizzare un sistema in grado di rilevare, in tempo reale, una serie di dati che possano essere elaborati in modo da ottenere importanti informazioni in merito ai valori di parametri fondamentali nella produzione di un vino. 

Below we present the details of the project that we have realized in this field and all the possible variations for which BOX-WINE can be used in other sectors of Smart Agrifood.

The entire project has been designed and developed to fit in as versatile a way as possible in the sustainable agro-food macro-trajectory, in particular, in the context of Big Data for the creation of complex algorithms and decision support systems for the production of DOC and DOCG wines.

The aim of the project is to create added value through the development of predictive models of wine shelf-life, using fast, economical, easy-to-use and smart methods of analysis. Therefore, as it is easy to understand the skills that such a system requires are certainly computer ones, in order to manage and interpret the big data generated through the use of device and artificial intelligence (IA)but we are also talking about oenological and analytic instrumental skills to interpret the information in the wine optics  

Oxidation is the main cause of the loss of freshness and aromatic elegance of the product, which undergoes a qualitative decay and the premature ageing of the wine. The oxidation phenomena are quite complex and see the involvement of several phenolic compounds including amino acids, SO2, peptides, aldehydes and various metals that combining together, under the effect of different levels of sun exposure and thermal changes cause a major change in the product’s characteristics. Generally, through a redox reaction catalysed by transition metals – Cu+ and fe2+ – the oxygen is converted into very reactive free radicals, capable of oxidising a large number of compounds present in the wine causing, so, the early ageing we were talking about before.

As a result of globalisation, the wine market has also suffered important consequences in the marketing of wines such as, for example, the fact that bottles often travel from one part of the world to another under conditions that are not adequately controlled. Moreover, to date, in the large distribution, where the rule of large numbers and just-in-time applies, these same bottles can remain exposed on the shelves for a period of time that cannot be established a priori, nor is it possible to tell when they will actually be consumed by the buyer. Therefore working on the improvement of shelf-life is difficult for the wine producer without adequate supports that help him to collect all the data and information he needs.   

White wines are particularly sensitive to oxidation: The varietal, fruity and fresh aroma is lost and replaced by heavy notes of honey and beeswax that are often accompanied by an incubation of color and the development of a bitter aftertaste. In red wines, on the other hand, appear aromas of dried plum and cooked fruit. On the palate the wine flattens and a darkening of the color is observed. Thus, with early ageing, both the varietal identity and different characteristics linked to the size of the terroir of that particular vine are lost, and the wines are perceived as very similar.

Control the solubilisation of oxygen in wine, avoid excessive exposure to air, eliminate metals (iron and copper) catalysts of oxidation, limit temperature changes that increase pH levels and reduce acidity by burning aromas are all necessary actions to ensure that the organoleptic qualities of the wine remain intact from when it leaves the farm to when the product is tasted and eaten it. 

The producer can act to prevent oxidation as long as the wine remains on the farm, thanks to the addition of additives, such as sulphur dioxide (SO2), which, however, being harmful to human health can only be used in limited quantities and, In any case, it is being used less and less. From the moment the bottle leaves the cellar it no longer has any control from the producer who, therefore, can no longer monitor any parameter related to the state of his wine.  

In this sense, understanding which wines and what particular chemical composition best support aging becomes a fundamental aid for the producer, as it allows to reduce the use of SO2 and other additives and to monitor the different parameters constantly with the possibility to act promptly to preserve the characteristics and qualities of the product. Therefore, controls on the effectiveness of treatments and production techniques are fundamental steps if you want to act in the improvement of the shelf-life of your wine by positioning it, so in the best way on the market.  

To date, the analysis of chemical markers, or profiling of metabolites to anticipate the shelf-life of a wine are often complicated, expensive, with rather long lead times, applicants for advanced equipment not present in wineries and consortia. In addition, data obtained from these approaches often require interpretation and evaluation by highly competent and trained staff who may be difficult to find

BOX-WINE: an intelligent unified system for vineyards and wineries’ control

The project that we are carrying out together with wineries of Soave has as objective the development of a predictive method of the shelf-life of the wine fast and economic with instruments that have been supplied to the producers and the consortium associates. Data acquisition units, specific sensors for experimental scenarios (such as Oxygen, Turbidity, Conductivity, pH, CO2, Sugar, Biosensors) and an ICT infrastructure for data storage and consultation are the technological components that we have used and that we have made available immediately.

In detail the predictive model aimed at improving the shelf-life of the product that we propose:

– adopt IoT cloud based solutions for data collection and aggregation in such a way as to obtain time series on the evolution of wine properties;

– time series analysis, using artificial intelligence techniques to be able to determine the qualitative decay and build its predictive models. 

We have also devised a monitoring kit specific to the wine sector that can collect the data of the parameters and that includes:

– Wireless sensors for microclimate control in wineries;

– Wireless sensors for the control of macroclimate in vineyards;

– Wireless control unit;

– Web/App application for remote control;

– Wireless sensors for counting, identification and access control.

We also provide training and refresher courses for the internalisation of skills related to machine learning with the aim of disseminating our knowledge from a computer point of view, ensuring that it joins that own oenologists and wine experts. In this way we are able to obtain a 360 degree training that allows those who work in the cellars to be able to make the most of the intelligent technologies of industry 4.0 that we provide them, interpreting the data with the right perspective. The systems of precision oenology and advanced process control that result from this project are easily expandable to other sectors of the agro-food sector. 

Therefore, what our project intends to promote is:

– develop knowledge of the oenological potential of electrochemical and spectrophotometric measures for wine classification;

– on the basis of untarget chemical data and sensory data, develop wine shelf-life classifiers through machine learning;

– implement classifiers for predicting the shelf-life of wines using IoT solutions and artificial intelligence; 

– to apply predictive models in wineries as a decision-making aid in improving production techniques to increase the shelf life of wines and to rationalise the use of SO2;

– apply predictive models as a decision-making support in the positioning on the market of the wines produced. 

IoT Solution for cellar monitoring

As for the use of BOX-IO in the wineries, what we can monitor is:

– Lightning: complete and total shelter from light is one of the most important factors for preserving the characteristics of a quality wine. The wine must be kept in the dark since the brightness activates and accelerates its evolution and oxidation processes. A wine kept exposed to the light for 6 months compared to the same wine kept in the dark will reveal a chromatic charge, a much more developed colour, a perfume and a taste much less intact and fragrant;

– Temperature: the temperature of the air is the second factor determining the time of evolution and development of bottled wine. The ideal temperature of the cellar is between 10 and 18 degrees, specifically around 12. High temperatures induce the acceleration of the evolution and aging of wine;

– Level of thermal inertia in the cellar: that is, the gradual but not sudden change in the heat inside the cellar during the seasons. It is the sudden changes in temperature (rising or lowering) that cause precipitation or early evolution of bottled wine. Less harmful a stable cellar temperature also around 20/22 degrees, than a sudden summer/winter variation from 12 to 24 degrees;

– Cellar air humidity: An excessively humid cellar induces on the one hand the proliferation of moulds on the upper outer wall of the cap, below the wine capsule; on the other hand the damage always from mould of the wine label; 

– Noise: it is precisely from noise that vibration is produced. A wine subjected to shock or vibration during storage evolves more rapidly than a firm one. The insulation of the basement walls, previously recommended for optimising temperature and humidity, is also very useful to soften and eliminate the possibility of noise vibration;

– Scent: it is important to avoid storing other aliments in the cellar, as the odours may contaminate the wine and make it unrecognisable.

– Ventilation: it is also essential to renew the air by avoiding sudden changes in temperature, by gradually following the recirculation.

Advantages of BOX-WINE on wine quality