*4.1. Cost Comparison between Additive Manufacturing Designs and Other Relevant Manufacturing Approaches*

A cost study has been carried out another time with the new design to compare the results with the original design and the cost of manufacturing it by plastic injection (see Table 8).


**Table 8.** AM costs assessment for the redesigned product.


**Table 8.** *Cont.*

If both designs are compared in terms of manufacturing times, a significant reduction can be observed and consequently production increases by a 41.4% per year. In addition, the reduction of the material used decreases by 72.5% compared to the original design. These factors directly impact part cost, decreasing up to 67.9%. Again, this cost is independent of the number of units manufactured. As seen before, injection molding reaches figures much lower of cost per part, providing that the number units to be manufactured is very high.

Figure 16 summarizes the cost levels of the three production strategies. Given the order level of 2000 units, the AM redesigned geometry is the production strategy that delivers the minimum manufacturing cost (12.86 € per part). The cutting production volumes between the different production strategies yield the different decision points. In the original AM design, AM production was the best preferred option for orders of 1045 parts or below, at a cost of 40.06 € per part. Otherwise, Injection molding would be the best desired strategy. In the part redesigned for AM, AM production is the best preferred option for orders of 3605 parts or below. For larger order figures, injection molding remains as the best production strategy.

**Figure 16.** Graphical comparison of the levels of manufacturing costs of the injection molding and AM strategies.

Figure 16 is thus the synthesis of the levels of cost for the current state of the art in AM and molding technologies. Providing that AM cheaper materials and faster AM machines can be developed, the horizontal cost lines for AM are expected to keep decreasing over time.

At the present time, the inventory taxonomy described in Section 2 shows that for inventory parts, normal ordering levels are very low. It has been found that it is very unlikely to require more than 2400 repetitions of a part in a year. Therefore, the zone of interest of the graph is below 2400 product units. From this selection, Figure 16 shows that substitution of manufacturing technologies—from conventional to AM—has economic sense only if DfAM is undertaken.

#### *4.2. Potential Impact in the Inventory Cost*

Following to the product filtering conducted in Section 2.1, the figures of the overall inventory parts considered in the present study yield a *TIMCConv*. *proc*. *<sup>α</sup>* calculated as formalised in Equation (1) of 1,350,670 €.

Then, in the case study, it has been possible to achieve a reduction in manufacturing costs per part of a 67.9% (*δ*) thanks to a reduction in the material utilised of an 72.5% in mass (*γ*) and a reduction on the processing time of manufacturing of a 41.4%. These percentages of reduction are aligned with the company previous experience and, according to the industrialists, can be considered a realistic average reference (not too optimistic or pessimistic).

Furthermore, given the company's data, it is possible to calculate the factors 'K' needed in Equation (6) to calculate de change in costs (ΔCs*i*) of the products analysed. The values of the cited constants can be found in Table 9.

**Table 9.** (Ki) constants associated to the different cost factors, calculated for the 2017 inventory figures at *Unistral Recambios*.


With these figures and considering that the original total mass of the target group for the switch to AM processes; i.e., fraction 'α' of the 'Plastics' category (m*Conv*. *proc*. *total <sup>α</sup>* ) is of 220,514 kg, the potential impact to the inventory associated costs can be found by utilizing Equation (12) as follows:

$$\begin{array}{ll} \text{\(\(\sum\_{i=1}^{i=1} \text{A}\)\text{S.s}\)} = \text{ } ^ {\text{Conv.}}\text{\(\text{Cour.}\)} \text{\(\(-0.679\) + (0.0333 + 0.0556 + 0.0222)\text{\}} \text{\(\(\text{Cour.}\)\)} \text{\(\(-0.725) + (-0.725) + (0.0488, 56)\)} \\ (-0.725) = -934,868.56,\end{array} \text{\(\(-0.725) = \)}$$

The overall reduction in costs is huge for the fraction 'α' of the product category of 'Plastics'. Indeed, it represents a reduction of a 69.22% of its Cs. Knowing that the entire TIMC of all the product categories ascended to 6,861,642 €, the possible cost reduction associated to the stocks change in manufacturing of fraction 'α' of 934,868.59 €, represents a 13.62% decrease in the inventory costs. This data yields a brief estimator of what can be feasible to achieve when switching from conventional manufacturing to AM technologies in the sector of the spare parts for fluid conduction systems.

#### *4.3. Considerations on Further Achievable Impact in the Inventory Cost*

As some considerations on the study and its hypotheses, it is important to mention that the present work completes a case study calculating a feasible cost reduction of a part—considered representative of the group, to estimate the impact that this exercise could have when extended to a relevant share of its product range (fraction 'α'). This estimation is performed on the context of switching from conventional manufacturing technologies—injection molding in the part analyzed—to AM technologies. Therefore, the study might be extended by selecting other parts in the inventory manufactured with different conventional technologies and so applying the methodology to yield additional cost reductions.

Also, the application of AM technologies could help achieving shorter delivery times; which could be incorporated in the study to quantify more precisely the cost reduction effects. So far, the implications that this would effectively have in the supply chain is out of the scope of the present study. Furthermore, as the batch size can be reduced to any figure, once switching to AM processes, the company might choose to modify the stocks level of the subset of products, this bringing further implications on possible savings. Moreover, some further plastic parts could be identified and redesigned from the category 'Kits', that could yield additional cost reductions on the overall stock cost.

These considerations on side-effects have not been analyzed in the present study to maintain a conservative savings quantity estimation. Future research may target the study of different inventory policies for the parts of the company, starting from the costing models formulated in the present article.
