The ILCIDAF project for the development of an Italian Life Cycle Inventory Database of agri-food products: the bread production phase

The Project of Significant National Interest (PRIN, 2017) entitled: "Italian Life Cycle Inventory Database of Agri-Food Products" (ILCIDAF), financed by the Ministry of University and Research, aims to promote the sustainability of the agri-food sector through the development of a database on a national and regional scale for some food chains that are significant for the Italian economy The objective is to create datasets for the Italian agri-food sector which current existing databases do not provide. Specifically, datasets that are geographically, temporally and technologically representative of the Italian national territory were developed. The selected supply chains are: bread and pasta, wine, olive oil and citrus fruits respectively studied by four scientific units: the Universities of Bari, Chieti-Pescara, Messina and Reggio Calabria respectively. The database is constructed considering the entire supply chain of the indicated foodstuffs. The University of Bari, to which the authors of this work belong to, has previously provided contributions on the agricultural phase of wheat, the milling phase and the pasta-making phase. In this work, on the other hand, the phases of the bread-making process of durum and soft wheat bread are described. The aim of this study is to construct datasets relating to the transformation phase of the indicated end products. In particular, 5 datasets were constructed: two relating to bread made from soft wheat flour, 2 with durum wheat semolina and 1 with wholemeal durum wheat semolina. The data was acquired through information provided by the companies in the sector (field data) through the completion of appropriately drafted questionnaires and company reports. In addition, the same analysed and processed were compared with data available in scientific literature.


Introduction
Several initiatives promoted and financed at the national level have been moving in recent years towards the creation of databases for the most representative sectors of the Italian economic system [1].These projects have the aim of overcoming one of the major critical points of the Life Cycle Assessment (LCA) methodology, which complies with ISO 14040:2021 [2] and ISO 14044:2021 [3], namely the lack of secondary databases [4] that are complete and representative from a geographic, temporal, and technological point of view of a given analyzed process.Current activities and projects include: 1) The Project of Significant National Interest (PRIN 2017) entitled "Promoting Agri Food Sustainability: Development of an Italian Life Cycle Inventory Database of Agri-Food Products" (ILCIDAF Project) financed by the Ministry of University and Research.2) The ARCADIA Project coordinated and developed by ENEA and financed with European Structural and Investment Funds (EIS funds) under the current Ministry of Environment and Energy Security.
3) The GRINS Project (Growing Resilient, Inclusive, and Sustainability) financed by the Next Generation EU program and with PNRR funds.
The ILCIADAF project aims to promote the sustainability of the agri-food sector through the development of a database on a national and regional scale for a number of food chains that are significant for the Italian economy (bread and pasta, wine, olive oil and citrus fruits).An in depth analysis of data concerning these food chains [5] highlighted that there are no representative datasets for the Italian territory.The research project consists of four scientific units: the Universities of Bari, Chieti-Pescara, Messina and Reggio Calabria respectively, each of which develops a specific food chain indicated above.The database is constructed considering the entire supply chain of the indicated foodstuffs.In particular, the Bari Unit, to which the authors of this work belong to, deals with the supply chain of pasta and bread, and the study was pursued by considering the durum and soft wheat cultivation phase, the milling phase, and the pasta and bread production phase as system boundaries.This last phase will be discussed in the paper, while the previous ones have been studied, analysed and published in other papers at other national conferences or in publications in indexed journals.

Materials and methods
The study envisaged the construction of datasets relating to the production phase of bread in different formats and using different types of flour (common, durum or wholemeal wheat).In order to achieve this objective, the following method was followed.The choice was made to analyse primary data, i.e. collected directly in production plants at major southern Italian companies (whose products are sold throughout the country) via questionnaires.The data were distinguished according to the type of product output: traditional bread made from durum wheat flour and wholemeal wheat flour, soft wheat bread for burgers and hot-dogs.The data acquired covered all inputs and outputs of the production plant.
The consumption of input raw materials necessary for bread dough production (quantity and type of flour input, water necessary for dough, eggs, oil, salt, yeast and other minor ingredients), the consumption of electrical and thermal energy necessary for the operation of the plant and for the baking phase of the bread, the materials for the packaging of the output product and the by-products and production waste were analysed.
In particular, the consumption of electricity and thermal energy, provided by the company energy bills and meters, is indicated in an aggregate manner.In the case of electricity from the grid, consumption per production line is distinguished.This detail made it possible to appropriately allocate consumption per output product knowing the quantity produced annually for each production line and type of bread.However, there is also the addition of electricity produced by a photovoltaic (PV) system which contributes to 11.6% of the total electricity consumption of the entire business process.This quantity of electricity is not differentiated for the various uses and for the various production lines, but simply guarantees a lower demand for energy from the electricity grid.
The total quantity produced with the photovoltaic system has been equally divided by all output products and in this way, the quantity of kWh per kg of product is the same between the different datasets.Furthermore, regarding the outputs, "Ingrao et al., 2018" [6] underlines that the weight loss (equal to approximately 1/3 of the dough) is due to the yield of the manufacturing process, i.e. the loss of mass that occurs naturally as a consequence of the cooking phase.To define these quantities, in order to identify the different outputs of the cooking process, it was chosen to multiply the factors provided by the company relating to the weight loss of the dough, by the water necessary for the dough, in such a way to determine the quantity of water evaporated following the cooking of the bread.
The remaining part is attributed to the material losses of the process: burnt dough, waste, residues due to baking, which will act as secondary raw material for the production of breadcrumbs by the same company.
Following this methodology, five datasets were identified for the different products of the plant.Subsequently, these results were compared with data collected from a literature review.A search was carried out by indicating the keywords necessary for the literature collection, i.e. bread, LCA and Life Cycle Assessment.
Of the available publications, those reporting data relating to the production process were selected in order to compare these data and the specific company data considering one kilogram of bread produced as a functional unit.A further comparison was carried out using the data published in EPDs (Environmental Products Declaration) relating to bakery products, although recognizing that the data published within the EPDs often report data that are too aggregated to be used for the construction of a database and for comparison with other much more specific and disaggregated data collected in the field [7].Finally, once the comparison was verified, the data were processed to build the ILCIDAF project database.The bibliographic search according to the indicated methodology led to the identification of 12 papers meeting the selection criteria (Table 1, for simplicity each paper has assigned an identifying letter).Of these, only 8 provide more or less complete data on the bread production process.Energy flows and greenhouses gases of EU (European Union) national breads using an LCA (Life Cycle Assessment) approach (e) 2017 [11] The contribution to climate change of the organic versus conventional wheat farming: A case study on the carbon footprint of wholemeal bread production in Italy (f) 2015 [12] Life cycle assessment of bread from several alternative food networks in Europe (g) 2014 [13] Product carbon footprint of rye bread (h) 2011 [14] The carbon footprint of bread The selected EPDs (CPT code 234, bakery products) follow the PCRs for bakery products and in particular EPDs for the production of soft, hard and wholemeal wheat bread were chosen with minor variations for minor ingredients.

Results and discussion
The results obtained have been included in tables so that a comparative analysis of them is possible.The datasets do not always analyse the same type of product, since some works are, for example, developed for local products which are characteristic of limited areas of the national territory, while others, on the other hand, are related to foreign products, or made using flour of a different nature or with the addition of cereals other than durum and soft wheat.The data collected on the farm were distinguished according to the output product and, in particular, Table 2 shows the 5 products (datasets) for these different types of bread leaving the plant.Table 3 reports all inputs and outputs related to raw material inputs, resource consumption, and output of products and by-products for the reviewed literature.
The eight selected reviewed publications allow us to identify 11 different types of products: seven breads made with soft wheat flour, one with semolina flour and three others with wholemeal flour.Of these, two products require the addition of rye and only one contains a small quantity of barley in addition to rye.
There were 5 EPDs selected, all related to the production of sliced bread from both soft and durum wheat with the addition of particular ingredients for some products.Table 4 shows the average values of the 5 EPDs for 13 similar products.The data collection carried out in the Southern Italian companies allowed the creation of inputs and outputs for 5 different types of bread made from both durum and common wheat.
The quantity of flour used per 1 kg of product is similar for the different types of bread, varying from a minimum of 0.651 kg (common wheat, hot-dog) to a maximum of 0.750 kg (durum wheat, Pagnotta di semola).However, bread made with common wheat flour tends more towards the minimum value, while that made with durum wheat requires a greater quantity of semolina.The amount of water for the dough is less for bread made with soft wheat flour.In fact, for soft wheat, the minimum amount of water used is 0.306 kg compared to a maximum amount of 0.311 kg, while semolina bread has an amount of water ranging from 0.490 kg to 0.559 kg.The consumption of electrical and thermal energy is slightly higher for the production of semolina bread than for bread made with type 0 flour.Gas consumption, however, is the same for both breads with type 0 flour (minimum value), while the maximum value is for the production of wholemeal durum wheat bread.Among production waste (output), the smallest quantity is for hot dogs, followed by muffoletta and wholemeal loaf.The highest values per kg of bread output are recorded for hamburgers and durum wheat semolina bread.These wastes should also include those that are naturally lost following the cooking process.Bread produced with semolina flour undergoes a greater weight loss in the oven than that of soft wheat due to the greater quantity of water required for the dough.The waste from baking durum wheat bread acts as a secondary raw material for the production of breadcrumbs.
Comparing these results with those extrapolated from the selected papers shows that the results obtained are comparable.Considering the average bibliographic results (Table 3) the dough for soft wheat bread requires 0.548 kg of flour, while for the durum wheat dough 0.724 kg of semolina is needed.The quantity of water needed for the dough varies between 0.25 L and 0.6 L for an average value of 0.52 L per kg of bread produced.Only for a particular type of bread (f), characterized by a mixture of almost 1/3 of barley and the remaining 2/3 of flour, a high quantity of water equal to 1.85 L is required and consequently more energy for the cooking.The average values of the secondary ingredients are also comparable with the quantities collected at the bakery.Considering, however, the consumption of electricity, the average bibliographic data are slightly higher.The average electricity consumption (excluding that associated with baking "f" bread as it is characterized by significantly different dough) is 0.28 kWh.The average energy consumption resulting from the bibliographic analysis is therefore higher than the average electricity consumption indicated in the system for all 5 types of bread, equal to approximately 0.14 kWh from the mains to which must be added the 0.03 kWh from photovoltaic for a total of approximately 0.17 kWh.
As regards the average data derived from EPD, they allow more than anything else to indicate an order of magnitude of the input and output quantities associated with the production of bread.
In fact, as indicated in [1] the data published in the EPDs are too aggregated for use in a detailed Italian database, without being able to associate fundamental metadata to understand the nature or provenance of a given quantity such as, for example, the type of energy, the co-products released, the emissions into the air.Furthermore, some quantities such as energy consumption or the quantity of water appear to be overestimated compared to the data identified at the plant and through the bibliographic review.In summary, the presentation of data in the EPD format does not allow us to have details on the recipes of the products represented.In fact, the data published in the EPDs are represented in a very aggregated manner and this makes comparison and comparison with other more detailed data very difficult.Furthermore, they appear to be insufficient for the development of a database, however their usefulness remains as they represent official documents published and associated with products widely consumed on the national territory.

Conclusions
The comparison of the data collected in the production plants of Southern Italy, of the literature data and of the average EPD data relating to Italian bakery products confirms the good campaign of collection of field data necessary for the construction of bread production datasets within the ILCIDAF database.However, there are small particularities due to the technology used, the type of energy, or particular types of flours used and minor ingredients which, in some cases, differentiate the compared products.
This emphasises the need for specific data as considerable differences are also due to the availability of certain ingredients, but also to the customs, traditions and production of products typical of a particular geographical area.The constructed datasets will subsequently be uploaded open-source to the ILCIDAF website.The methodology carried out for the reprocessing of the plant data allowed the creation of 5 datasets relating to bread-making: 2 relating to bread produced with type 0 flour for burgers and hot dogs, 1 dedicated to a bread produced with durum wheat semolina of small dimensions and another 2 relating to loaves (larger in size) one produced with durum wheat semolina and the other with wholemeal semolina.From the results obtained, we observe a lower need for input raw materials (flours, minor ingredients and water) for bread produced with type 00 soft wheat flour, while the quantities required for bread produced with semolina are greater.The same observation is valid for the consumption of electricity, remembering that of the total consumption of electricity consumed, that from the network is appropriately divided by production line (88.4% of the total), while that produced by the photovoltaic system, equal to 11.6% of the total is provided in an aggregate manner without being able to distinguish the type of use (i.e. for production lines or for other uses).In addition, it would be desirable for more companies in the sector to collaborate in initiatives and research activities aimed at the construction of specific and representative databases for the bread production phase through the collection of primary data in the field.

Table 1 .
Bibliographic analysis: list of selected works.

Table 2 .
On-field data at bread production plant in Southern Italy.

Table 3 .
Data obtained from bibliographic review