2.1.5. Allocations

The scenarios under study are multi-outputs processes in which the managemen<sup>t</sup> of MSW is the main function of the system and the production of electricity and compost are additional functions. To handle this problem ISO [31] establishes a specific allocation procedure in which system expansion is the first option. In this case, in the anaerobic digestion stage, methane is assumed to be combusted with a 25% e fficiency of the low heating value of the biogas to generate electricity. The delivering residue of the anaerobic digestion, i.e., digestate, is transferred to a composting plant for the production of biocompost. While the compost is assumed to replace mineral fertilizer, with a substitution ratio of 20 kg N equivalent per ton of compost, the energy generated from biogas is considered to replace the Spanish electric mix [47].

#### *2.2. Life Cycle Inventory (LCI)*

LCI consists on the collection of the relevant input and output data for the assessed systems [48], and is one of the most e ffort-consuming phases of an LCA, both in terms of work and dedicated time [44]. In this section, data sources and main assumptions for the scenarios under study are detailed. In this case, the LCI was collected by means of an input-output analysis. Regarding the processes shared by the two scenarios, such as the sorting plant and the composting plant, data were obtained from Cobo et al. [46] and Righi et al. [49], respectively. Since the amount of MSW entering pre-treatment and treatment stages depends on the scenario analysed, the data of the sorting plant and composting plant collected in Table 2 are expressed per tonne of waste treated.


**Table 2.** Life cycle inventory for door-to-door collection and pneumatic collection systems.

#### 2.2.1. Door-to-Door Collection

Primary data were obtained from the Spanish Ministry of Environment and Rural and Marine Affairs (MERMA) [50] and from the UNESCO Chair in Life Cycle and Climate Change (by means of a personal communication from Ecoembes in the framework of the Life-FENIX project), while secondary data were sourced from the thinkstep database [51]. The fuel consumption rates were estimated at 3.98 and 0.07 L of diesel and gasoline per collected t of waste [50]. Energy demand and emission factors for fuel production were taken from the thinkstep database [51]. Based on Ecoembes information, containers requirements were estimated at 0.107 m<sup>3</sup> per FU. All fractions were assumed to be deposited in a high density polyethylene (HDPE) container, which was modelled assuming the following composition: 74.75% HDPE, 21.57% steel and 3.68% styrene-butadiene-rubber. A lifetime of 7.5 years was assumed for all the containers. They were considered to be washed six times/y, using 0.35 kg of detergent per m<sup>3</sup> of container. The LCI of the system is shown in Table 2.

#### 2.2.2. Pneumatic Collection

The main sources of information were Ecoembes and the two main companies that manage the pneumatic system in Portugal and Spain: Envac Iberia [52] and Ros Roca [53]. The authors personally visited their plants aiming to have a better perception of the system performance and to ge<sup>t</sup> real and representative data allowing the construction of the model as close as possible to reality. Secondary data, such as the production of stainless steel cold rolled for the manufacture of bins and pipes, were taken from the Thinkstep database [51]. Data collection is explained according to three main infrastructure stages of the system: waste collection plant, central collection points and underground pipes. The LCI data is listed in Table 2.

• Waste collection plant

All the waste gathered in the collection points is carried to the waste collection plant through the network of underground pipes. Then, waste is compacted and deposited in a container until it is full. The energy consumed by the system is linked to the suction and compaction operations at the waste collection plant. The suction stage is estimated to work 3,500 h/y with a power of 220 kW. The average compaction demand is established at 15 kW for 1,100 h/y. From these data, the consumptions per metric ton of waste collected for both processes has been calculated, obtaining the total consumption of this waste collection system. The estimation is explained in detail in Section S1 of the Supplementary Material.

• Central collection points

The central collection points are composed of a number of bins where citizens can deposit the wastes. Below these boxes, there are the waste valve and an air valve which are responsible for connecting the bin of the central collection point to the overall network of underground pipes. These components have been excluded from the system because of their irrelevance comparing to the whole environmental impact of the system. According to Envac Iberia [52] and Ros Roca [53], the number of bins per collection point depends on the amount of waste fractions managed. In the model under study, it has been considered only one fraction at time (the organic fraction). A lifetime of 9 years is assumed for the bins [54]. Two different containers composed the system: a sidewalk container, entirely made of stainless steel, and an indoor bin, made of stainless steel and a small fraction of fiberglass and aluminium. However, both containers have been assumed to be made entirely of stainless steel because the contribution of fiberglass and aluminium is negligible compared to the stainless steel. According to Ecoembes [28], there are 70 collection points per kilometre and, in this case, one bin per collection point having a weight of 40 kg each [55]. In Section S2 of the Supplementary Material the procedure for the calculation of the weight of bins is described.

#### • Underground pipes

The underground network of pipes is made of stainless steel pipes with an average length of 1 km per station line and an inner diameter of 50 cm and 12.5 mm of thickness. A 30 years lifetime is considered for them. Inside of these pipes, a stream of air transports waste bags at an average speed of 25 m/s. [55]. In Section S3 of the Supplementary Material the weight of pipe per tonne of waste collected is estimated.

#### **3. Results and Discussion**

#### *3.1. Comparison of the Two Waste Collection Systems*

The four scenarios were assessed following the LCA methodology. Figure 5 shows the primary energy demand (PED) results per FU, also cited as cumulative energy demand (CED), an indicator commonly used to assess waste managemen<sup>t</sup> systems [56]. The results are divided into the four stages of the life cycle (collection, transport, pre-treatment and treatment). As it can be observed, the results change significantly when the organic fraction is collected separately (scenarios 1b and 2b) compared to scenarios 1a and 2a where the bulky fraction includes the biodegradable fraction. In this case, the PED of scenarios 1b and 2b is similar: −3,128 MJ and −2,701 MJ, respectively. The negative values are associated to the energy savings of the anaerobic digestion process. This is because the environmental benefits of electricity and compost production displace the environmental impact of production of electricity from the Spanish grid mix of 2016 and the production of fertilizers, and overcomes the energy inputs of the collection system. It is important to correctly calculate the credits obtained through material recycling and energy recovery [32]. Those results indicate that, when the biowaste is collected separately and, considering the composting of these residues, the use of a pneumatic system could be a suitable option, as well as the door-to-door collection was. However, when the bulky fraction includes the biodegradable fraction, pneumatic collection (scenario 2a) exhibits the largest PED, estimated at 405 MJ per FU, whereas the door-to-door collection presents negative PED values (−245 MJ). The reason is that the energy demand for the vacuum system is higher than the energy recovered from the composting process. Authors such as Iriarte et al. [30] and Punkkinen et al. [38] stated that the system with the greatest environmental impact is the pneumatic collection compared to door-to-door and multi-container collection. Nevertheless, these studies only include the collection stage, without the subsequent waste valorisation treatment and environmental benefits. On the other hand, considering the CED indicator, Iriarte et al. [30] stated that the door-to-door system had the greatest energy demand, in particular 38% higher than the pneumatic system. In our study, this di fference between door-to-door and pneumatic collection is lower, around 14%, due to the energy requirement for the anaerobic process. Finally, the contribution of the MSW transportation to the pre-treatment installation and the contribution of the pre- treatment in the sorting plant was negligible, around 1% (in case of scenarios 1b and 2b the pre-treatment stage is not necessary, therefore, its contribution to the total PED is zero). These results reinforce the idea that the use of the pneumatic system could be an appropriate waste collection system in the development and the implementation of smart city technologies in historic areas of cities [57]. As mentioned previously, o fficial data sets that street-side containers is the most common MSW collection system used. In addition, some authors such as Iriarte et al. [30] and Aranda-Usón et al. [39] have stated that this collection system presents the lowest environmental impact. However, since the main objective of this study is assessing the managemen<sup>t</sup> of the organic fraction generated in historic areas of cities where the conventional street-side collection systems is not feasible due to tra ffic restrictions, this type of collection was excluded from the study. In this sense, there is a debate about which is the best collection system for these areas. Currently, there are not studies that address this problem from an environmental point of view and, in particular, under an energy and climate change perspective. In this sense, historic areas of cities are suitable zones to apply smart city technologies where monitoring and information are essential for the correct managemen<sup>t</sup> of these areas. The managemen<sup>t</sup> of MSW is an example for this purpose.

**Figure 5.** Primary energy demand (PED) per functional unit of each scenario.

Figure 6 focuses on the collection system stage and displays the PED consumption of each process of the waste collection system (manufacturing of components, maintenance and operation). As can be observed in Figure 6, when the waste collection system is considered isolated, the pneumatic collection consumes 5.0 times more primary energy than the door-to-door collection. This is in agreemen<sup>t</sup> with Iriarte et al. [30] and Punkkinen et al. [38] and highlights the importance of considering the subsequent waste valorisation treatment and environmental benefits. Regarding the contribution of the different processes, in the conventional door-to-door collection (scenarios 1a and 1b), fuel production for garbage trucks in the operation step was responsible for 64% of the PED. The rest of PED is attributed to the production and maintenance of containers (36%). On the other hand, in the pneumatic system, the production of electricity used in the process accounts for almost 100% of the total PED, whereas the manufacturing of bins and pipelines and its maintenance is negligible compared to the consumption of energy.

**Figure 6.** Primary energy demand (PED) of each process of the MSW collection system.

According to the results of the study, the electricity requirements for vacuum production is the item with the most impact. Therefore, reducing energy consumption for waste transport through the underground system and using a more environmentally friendly energy source, such as renewable energies [58], are the best improvement measures to ensure the sustainability of the pneumatic process, since waste managemen<sup>t</sup> and energy supply systems are becoming more inter-connected [59]. This study contributes to decision-making in waste managemen<sup>t</sup> strategies and enables the introduction of the environmental variable in the design model of smart cities. In addition, this study comprises the basis for the future optimisation of pneumatic collection as hybrid systems feed by renewable energies, such as photovoltaic installations.

#### *3.2. Sensitivity Analysis*

Our results in the baseline case study depend on many assumptions concerning the installation of the pneumatic system. Particularly, the effectiveness of the biodegradable collection, the number of waste collection points in the pneumatic system and the population density were examined.

#### 3.2.1. Effectiveness of the Biodegradable Collection

The energy efficiency of the pneumatic collection system depends on the attitude of the citizens towards the introduction of the selective collection of the organic fraction. In this sense, a sensitivity analysis was performed varying the effectiveness of the biodegradable collection, considering the best scenarios (scenarios 1b and 2b), in which 100% of the organic waste generated is collected separately, together with less-efficient scenarios in which part of the organic fraction is collected separately and the rest remains in the bulky fraction. Figure 7 shows the results of the sensitivity analysis.

**Figure 7.** Results of the sensitivity analysis varying the effectiveness of the selective collection of the organic waste (100%, 80%, 60% and 40%).

As previously mentioned, the negative values are associated with the energy savings from the composting process. When the efficiency varies from 100% to 40%, the energy savings are reduced to 57% for door-to-door collection and to 66% for the pneumatic system. Therefore, these results highlight the importance of people awareness and information campaigns about the benefits of the biodegradable waste being recycled and aim at emphasizing the role that consumers play in biowaste separation at source [17].

#### 3.2.2. Population Density

Taking as reference 20,000 citizens per km2, to see how the environmental impact is influenced by higher densities and, therefore, larger waste volumes, the number of citizens was increased up to 80,000 citizens per km2. Given the fixed capacity of the collection system, the PED will be established by the increase in the emptying frequency due to the increase in the waste volumes. For the sake of comparison, the managemen<sup>t</sup> of higher waste volumes in the door-to-door collection requires either a fourfold increase in waste containers, a four times higher emptying frequency, or an intermediate solution. In this case, a higher emptying frequency is considered [27].

The increase in the population density in a specific zone is mainly due to tourism. In Spain, cities such as Fuengirola (Málaga), Benidorm (Alicante), Ibiza (Balearic Islands) and Salou (Tarragona) experiment an increase of between 65–75% in their population density due to tourism [60]. The tourism can sustain high levels of employment and incomes in the economies of many regions, but the sector is a source of environmental impacts, resource consumption and public health problems [61]. In particular, one of the most important impacts of tourism is the generation of MSW. According to Mateu-Sbert et al. [61], on average, an increase of 1% in the tourist population in Menorca causes an overall MSW raise of 0.282%. Following this statement and considering that a population density of 20,000 citizens/km<sup>2</sup> generate around 1,936 metric tons of MSW per year [27], the increase of the waste volume due to the raise of the population density was estimated. On the other hand, the composition of generated MSW is extremely variable as a consequence of seasonal and lifestyle impacts, in particular in historic areas of cities, which are a tourist attraction [62]. In this sense, it is considered that the increase in the population density a ffects the waste composition. A population density of 20,000 citizens/km<sup>2</sup> in winter, could corresponds to 40,000 citizens/km<sup>2</sup> in autumn period, reaching 60,000 and 80,000 citizens/km<sup>2</sup> in spring and summer, respectively. Figure 8 displays the di fferent waste generation and composition for each population density considered.

**Figure 8.** Variation in the municipal solid waste (MSW) generation and composition due to the increase in the population density.

The amount of MSW increases from 1,936 metric tons with 20,000 citizens per km<sup>2</sup> to 2,586 metric tons when the population density reaches 80,000 citizens per km2. Regarding the waste composition, there is no a grea<sup>t</sup> di fference in the organic content, varying from 36% in winter to 41% in spring. These results were introduced into the environmental model previously described to analyse the model sensitivity to changes in the population density. Figure 9 displays the results of PED obtained for door-to-door collection and pneumatic collection of the biodegradable fraction. It can be observed that the higher increase in the population density, the greater negative value of the PED, which means a higher recovery of energy from the treatment of the MSW. An increase of 75% of the population density means 27% of PED savings for both collection systems. These results indicate that, even though the increase of MSW generation implies a higher emptying frequency for the door-to-door collection, and a higher use of energy for the pneumatic collection, since more daily pneumatic transportation is needed, the savings from the anaerobic digestion treatment are higher.

**Figure 9.** Results of the sensitivity analysis varying the population density of the area considered.
