*3.2. Case Study (Spare Part for Fluid Handling): Product Redesign and Virtual Validation*

### 3.2.1. Preliminary Infill Study

As a starting point in the product redesign it is interesting to evaluate the effect of the reduction of the infill percentage of the part. Providing that the external features of the case study part remain the same, the same model can be printed with a different internal structure (filler or infill). The infill modification as a percentage is a very powerful approach, due to its easiness, to reduce the amount of material needed while maintaining the necessary stress resistance of the parts. In many cases, the filling of the part does not need to be 75% to 100%. A 25% fill can also provide enough strength and saves time and print material.

Also, in FFF, very solid parts can yield to warping and deformations due to the cooling process of the part. This deformation or warping can be reduced either through good adhesion to the construction platform or also through the printing configuration. In particular, it can be significantly reduced if a less dense structure is used.

One of the best approaches for tuning the infill level is to test it live by printing a part. Figure 6 shows a couple of 3D printed repetitions of the case study part. The part in the left was manufactured with an infill level of 25%, while the one in the right was manufactured with an infill level of 75%. Only with this infill percentage reduction, a significant change in the shape of the part the material is reached. In particular, for the printed tests, the shape distortion is reduced between 5 and 7 mm in the four lower edges of the part. Furthermore, a significant cost reduction is observed as the printing time is reduced by a 30%.

These exploratory results can be utilized as preliminary quantitative figures for a starting point of the specific product improvement. Based on these, the product redesign, in form of DfAM, should be explored in detail undertaking the complete topological optimization and analysis.

**Figure 6.** Case study printed parts: (**a**) Support manufactured with an infill percentage of 25%; (**b**) support manufactured with an infill percentage of 75%.

#### 3.2.2. Product Redesign and Virtual Validation

The computer aided design (CAE) analysis starts with the definition of the constraints imposed by the components in the context of the system. As such, the first element to be considered for the simulation is the gold membrane that exerts a pressure on the surroundings of the housing of the support (see Figure 7a). The central part of the membrane also exerts pressure, but only when there is water pressure inside, for this reason it is presented in the study of loads. The material of the membrane is EPDM rubber (ethylene propylene diene monomer rubber), a rubber derivative with 6 mm thickness and a modulus of elasticity of 10 MPa.

The support object of the case study in the simulation is shown in Figure 7a represented in blue color. However, for the redesign and simulation means, the support is divided into two domains. The outer shell or skin, for aesthetic reasons, should remain as it is in the original design, but the interior of the support can be modified. Therefore, as presented in Figure 7b the part is divided into the design domain (interior, in red) and the non-design domain (housing or skin, in blue) with a thickness of 2 mm.

**Figure 7.** Case study part: Support (in blue) and membrane (in gold). (**a**) Support and membrane; (**b**) design domain (red) and non-design domain (blue).

Secondly, the study of the supporting surfaces of the part is addressed. The biggest surface of part is in direct contact with the base of the metallic piston and therefore, there must be included a surface restriction (see Figure 8a). As for the screws, which joins the body with the piston, it restricts movement in the X, Y and Z directions (see Figure 8b). In the center of the part there is a through housing where the rod and the actuator pass through the support until they encounter the membrane. The loads on the sides of the housing are not considered in this case study, assuming the actuator next to the stem will not contact with the walls around it.

**Figure 8.** Case study part (support, in blue) and membrane (in yellow): (**a**) Surface (red) including a restriction; and (**b**) four screw fasteners support (yellow), which are the points at where the model does not allow any physical displacement over any of the cartesian directions (X-Y-Z).

The third step is to identify the loads in the model and two extreme conditions have been defined: Valve totally open, and valve totally closed. In the first condition, the pneumatic actuator is resting, and the membrane passes the flow through the interior of the body at a nominal pressure of 10 bar (PN10). In this case, the pressure acts on both sides of the support (see Figure 9a). In the second condition, the pneumatic actuator applies force on the membrane stepping on the body and preventing the passage of the flow inside. In this case, the pressure acts only on one of the faces (See Figure 9b).

The body and the support are assembled by means of screws with a tightening of 6 Nm, in both cases a pressure of 1 MPa of compression has been applied on the membrane that corresponds to a deformation of 0.6 mm in its dimensions. The *Novamid® ID1070* material (which is the filament used in the manufacturing of the present case study) was characterized for the two most common printing patterns (0-90 and 45-45) for the X-Y printing plane (see Table 6).

**Figure 9.** Case study part (support, in blue) with load in the conditions of: (**a**) Valve totally open; and (**b**) valve totally closed.

**Table 6.** Mechanical properties of *Novamid® ID1070.*


From this material characterization, the virtual testing has been executed assuming a linear simulation and utilizing the least favorable values for the material properties that are presented in Table 7 (X-Y print at 0-90). A safety factor of 30 has been considered in the voltage factor for the simulation.


**Table 7.** Mechanical properties imposed in the virtual testing conditions.

Once the model is ready, the first simulation run is used to evaluate the behavior of the initial totally solid (100% level of infill) model of the part. With the simulation results it is possible to quantify to the material deformation (as presented in Figure 10a) as well as the Von Mises stress values (as presented in Figure 10b), providing detailed information on the different model areas.

**Figure 10.** Case study part (support): (**a**) Displacements of the material for the solid model; and (**b**) Von Mises stress for the solid model.

In particular, the maximum displacement results are 0.031 mm obtained at the top of two of the sides of the part. The maximum stress of Von Mises is estimated around fixations with a value of 3.02 MPa (see Figures 10b and 11a).

**Figure 11.** Case study part (support): (**a**) Von Mises stress focused on fixations; and (**b**) results of topological optimization.

These values obtained in the simulation indicate that the model is oversized and that there is room for material optimization. In this case, the objective of the topological assessment is to propose material to remove in those points where the interaction of the loads is low, and/or no displacements occur. In this manner, the objective function is set to maximize the rigidity of the model, taking into account the restrictions and supports. The results of the topological optimization are presented in Figure 11b.

The results obtained in the analysis indicate the areas where it is necessary to have a greater density of material so that the part is functional according the specified mechanical requirements. Due to

the asymmetry in the load results, the proposed design in this case is not completely symmetrical. However, to make sure that the part would meet the solicitations in any of the possible mounting dispositions, the denser side has been mirrored to the opposite side of the initial plane of symmetry of the part. Once the redesigned model is completed, a new simulation is run to evaluate the level of displacements and stress. The results are presented in Figure 12a,b.

**Figure 12.** CAE of the case study part (support): (**a**) Displacement results in the optimized design, and (**b**) Von Mises stress result in the optimized design.

In the redesigned product, the maximum displacements meet are of 0.138 mm, which are considered acceptable within the normal operation limits. Furthermore, the maximum stress (mostly due to compression efforts) is of 26.2 MPa, falling on the safe operating area.

3.2.3. Manufacturing Specifications for the Product Redesign

To achieve the desired dimensions in the printed part, thermal contractions must be considered. For this, it is necessary to over-scale the three-dimensional model, increasing the volume in a proportion of a certain value. This value is dependent on the printing conditions, but in general, for *Novamid® ID1070 1.02*, a 2% of increase in dimensions can be used. In this way, it is possible to correct the deformation contraction effect by ensuring the correct dimensions after printing (see Figure 13).

**Figure 13.** Original model (**left**). Corrected model (**right**).

*Novamid® ID1070* is printable with FFF machines whose extrusion nozzle head with temperature ranges between 220–245 ◦C. The most usual printing temperature is 230 ◦C, achieving the homogeneous melting of the material at temperatures above 225 ◦C. The optimum mechanical properties are observed at temperatures between 225–245 ◦C.

The *Novamid® ID series* is a range of high performance filaments for extreme resistance and ductility. These properties are closely related to the level of crystallization of the material. In order to achieve performance similar to the standard injection molding of PA6, the high level of crystallinity in the material has been maintained. Because of this increased crystallinity, using the optimum printing conditions is necessary.

To print the new models, it was used with the BCN3D Technologies Sigma 3D with the same printing values for the parameters of nozzle diameter, filament diameter, print speed, layer height, extrusion temperature and temperature of the hot bed than in the initial printing. The infill parameters are set at 100% in this new trial of manufacturing. The reason is that once the part has undergone the topological optimization, the material allocation needs to match the computer aided design.

Concerning the adhesion of the printing part to the construction plate, the accumulation of stresses due to contractions during printing can cause the separation of the printing substrate. To this regard, adhesion can be increased by chemical bonding or mechanical bonding. To achieve a correct adhesion in the hot bed, an adhesive promoter (*Dimafix®*) was used, also adding an edge to increase the contact surface.

#### 3.2.4. Physical Functional Validation of the Redesigned Printed Part

After the virtual Computer aided engineering testing, the printed parts have been functionally validated in a specific test bench at the company *Fluidra*.

The test bench consists on a small circuit composed by a compressor that introduces pressurized atmospheric air into an accumulator that is presented in Figure 14. A pressure regulator set at 10 bar allows the air flow through a normally closed solenoid, which commands the piston that lifts the membrane. As presented previously in Section 3.1.1, the movement of the membrane (opening and closing) allows and stops the flow of liquid through the main body of the valve. Therefore, the support analyzed in the present case study is mounted between the piston and the main body of the valve. The solenoid valve is controlled by a square signal, with short cycles of five seconds with voltage and five seconds without voltage. According to the established protocol, the study part (support) must meet the same minimum operating life as the minimum membrane life, which is set at 5000 cycles.

**Figure 14.** Schematics of the testing set for the pneumatic valve according to the protocol at the *Fluidra* group. Contains the solenoid valve, the piston and the compressor.

For the validation of the present study, three printed samples were tested during 5500 cycles to check the resistance of the support in accelerated operation. All three sample parts were able to be mounted correctly, fitting in size and not having dimensional or tolerance problems. Furthermore, the three parts worked correctly showing no damage or defect in their geometry. The physical bench set is depicted in Figure 15.

**Figure 15.** Physical testing of the printed part. It can be seen how the command circuit is mounted on a main fluid circuit. (**a**) Detailed view of the support mounted under the pneumatic piston. On top, the solenoid valve. (**b**) General overview of the fluid testing set, containing the compressor, valves, piston and the case study support.
