Optimization approach of DED process to fulfil the requirements on material properties and component performance of waterjet impeller

Directed energy deposition (DED) processes have a great potential for ship building industry to reduce the lead time of ship part supply. However, ship parts such as propellers and impellers have critical requirements on material’s properties and component’s performance. To ensure the exploitation of DED manufacturing for critical parts, an optimization approach of DED process was developed and demonstrated on waterjet impeller made of duplex stainless steel. The developed approach is based on the understanding of the effects of DED process and heat treatment on the material microstructure, and the effects of the microstructure on the material properties, all optimized to fulfil the requirements on the component. The manufacturing and qualification testing were planned based on the requirements in accordance with the current standard guidance. The assessments by standard testing methods, including tensile, Charpy impact, and bending tests demonstrates that the DED material with an optimized heat treatment fulfil the standard requirements. Corrosion and cavitation tests were also performed and demonstrates the high resistances of DED material to cavitation erosion.


Scope
Critical components in the propulsion and pump systems, such as propellers and impellers, are mostly produced by sand casting.Sand casting is a proven technology for different metals and sizes, but the major disadvantage is the complexity of the process and the long casting cycle involving, patternmaking, core making, molding, melting, pouring, and finishing [1].Additive Manufacturing by Directed Energy Deposition (DED) represents a promising alternative to sand casting, particularly for customized production [2].For on-demand manufacturing, DED enables significant reduction of the production time due to the reduced manufacturing steps and the very high build-up rates.It allows empowering the design of maritime parts by enabling complex geometry and advanced multi-material combination that can't be produced by conventional manufacturing.
Although the material properties of parts produced by DED are in many cases comparable with the material properties of parts produced by conventional manufacturing [2][3] [4], there is still a critical need to understand the effect of the DED process on material properties to improve the quality of DED manufactured parts and to prevent defects such as anisotropy, porosity, lack of fusion, voids, poor surface, part distortion, and residual stresses.The aims are to optimize the control and increase the reliability of the DED process and guarantee the quality requirements of manufactured parts.
In previous work investigating DED manufacturing of duplex stainless steel [7], the deposition speed was found to have a significant effect on the microstructure characteristics of the austenite and ferrite phases.Such effects were found to affect the mechanical properties represented by the measurements of hardness, tensile stresses, tensile elongation, bending force, and impact energy.While a balanced ferriteaustenite microstructure was obtained for different deposition speeds, the material produced with increased deposition speed demonstrated higher strength, but more brittle behavior [7].The reduction of deposition speed promotes the formation of coarse microstructure which contributes to improving the ductility.However, the produced materials failed to fulfill the requirements on tensile elongation and impact energy.To fulfill such requirements, the current work investigates a post-processing heat treatment (HT).It targets improving the microstructure morphology of the phases to achieve the required properties.The DED and HT were optimized and evaluated according to current standard recommendations for cast materials [5][6].Testing of the mechanical properties after a build process qualification and part qualification were performed and compared.In addition to the assessment of material properties, microstructure characterization was performed to provide experimental databases to support the understanding of the effect of DED process and HT post-process on the microstructure and the material properties.

Manufacturing optimization approach
Figure 1 illustrates the manufacturing approach, where the requirements for the waterjet impeller provide the basis for the established manufacturing and qualification testing plan, in agreement with the current standard guidance.The optimization approach for manufacturing is based on the understanding of the dependencies between process parameters, material microstructure and material properties.Standard tests and microstructure characterizations were performed to support the optimization approach, in which the process parameters target an optimized microstructure for the required material properties.

Requirements
According to EN ISO 17296-3 [5], the specifications of the main quality characteristics of the parts and the related testing methods depend on the level of the part criticality.The waterjet impeller is defined as a critical part, where the corresponding additive manufacturing category (AMC) is designated by AMC3 (Ref.DNV Class Guidance [6]).Hence, the part shall be built using a qualified building process, and part qualification testing, in addition to production testing, shall be performed on the manufactured component [6].The requirements on the geometry and tolerances were defined in a CAD file transferred to the AM machine.The material chosen for the component was 2205 duplex stainless steel, for which the composition is defined according to UNS S31803 and can be found in [6].An important requirement on the material microstructure is the balanced ferrite-austenite microstructure, where the fraction of ferrite in the material must be between 40% − 60%.The microstructure must not include any intermetallic phases.
The requirements on the mechanical properties are listed in Table 1.Standard testing methods must be applied to assess the mechanical properties [5][6].In addition, material tests including hardness, bending and cavitation need to be performed.The properties measured by these tests will be compared to the properties measured for equivalent cast material.

Manufacturing and qualification testing
The DED manufacturing was performed using a laser metal deposition machine (Trumpf TruLaser Cell 3000) equipped with a 3000 W fiber laser.Argon was used as a shielding gas.The heat treatment was performed in a vacuum furnace according to standard recommendations to avoid oxidation with nitrogen gas for the cooling.The manufacturing plan involves three steps associated with building process qualification (BPQ), part qualification (PQ), and a production qualification, in accordance with standard recommendations [5][6] (see Figure 2).First, a small-scale testing was performed to identify the preliminary values of DED process parameters which allowed the production of material with the balanced ferrite-austenite microstructure.Single bead specimens of 30x30 mm 2 and S-shaped specimens with variable thicknesses were also considered.Then, a manufacturing specimen in the form of 75x75x130 mm 3 block was used to extract the specimens required for material testing and microstructure characterization during BPQ.The evaluation of the effects of the actual geometry and section transitions was then considered in the PQ step.The geometry of the manufactured PQ specimen represents the geometry and section transitions which would be found in the waterjet impeller (see Figure 2).The qualification testing in the PQ step was equal to the testing in the BPQ step, with additional tests performed to evaluate the component performance (e.g., geometrical tolerance, cavitation erosion resistance and fatigue properties).The last step is the production of the final waterjet impeller and the necessary qualification.The quality checking at this step includes testing such as geometrical and surface checking and penetrant testing.Limited tensile and Charpy impact tests are performed on an extra-specimen produced together with the impeller.
In the current paper, material tests performed in the PBQ and PQ will be presented.More details on the manufacturing plan and qualification testing plan, including BPQ, PQ and production testing will be presented in an extended paper.

DED and heat treatment
The optimization of DED process leading to a balanced ferrite-austenite microstructure was reported in a previous paper [7].The optimized values are listed in Table 2.The main process parameter influencing the microstructure and the mechanical properties was found to be the deposition speed.An increased value of the deposition speed resulted in reduced ductility due to the effects of the deposition speed on the microstructure morphology as discussed in [7].While an optimized value of the deposition speed was defined to fulfill the requirements on yielding stress and ultimate stress, the obtained tensile elongation and Charpy impact energy were found to be below the requirements.To fulfil the requirements on the tensile elongation and Charpy impact energy, a postprocessing heat treatment is applied to modify the microstructure morphology of the phases and improve the ductility.
The applied HT process consists of heating the specimens to 1100°C, holding the temperature for 2 hours, and rapid cooling in nitrogen gas.To ensure consistent heat treatment, the BPQ specimen was divided into four sections representative of the thickness of the impeller, and the temperature was controlled using thermocouples positioned in the middle of the sections.For the PQ, thermal sensors were installed in the thick and thin sections of PQ specimen to measure the heating and cooling rates and to evaluate the effects of section transition on the heat treatment.The microstructures and mechanical properties of heat treated PBQ and PQ specimens are compared to the microstructure and mechanical properties of BPQ specimen as built.The results discuss the combined effect of heat treatment and part geometry on the material properties as required for waterjet impeller.

Material testing methods
The ferrite fraction was determined using a Feritscope Fisher FMP30, quantifying the ferrite content based on a magnetic measurement on fine ground surfaces, obtaining at least 10 measurements on specimens extracted from different positions in BPQ and PQ specimens.The tensile test, impact Charpy test and bending test were performed according to the standard procedures defined by ISO 6892-1 [8], ISO 148-1 [9], and ISO 7438 [10] respectively.Figure 3 shows the geometry of the testing specimens.Note that the tensile specimen in PQ was slightly shorter than for BPQ due to space limitations.To evaluate the effects of the building and deposition directions on material anisotropy, three specimens were extracted from three directions X, Y and Z, where X and Y are in the building plane and Z is in the building direction (perpendicular to the deposited layer).The specimens for microstructural analysis and hardness measurements were extracted from different positions in the BPQ and PQ specimens.The specimens were grounded and polished according to standard metallographic procedures.They were etched using Beraha etch [13] to color the ferrite and highlight the contrast with the austenite.A microstructure examination using optical microscopy was performed to document the microstructure morphology and to check if any defects and porosity were present in the structure.The hardness was measured by Brinell hardness measurement, HBW2.5/187.5.The cavitation erosion resistance test was performed according to the procedure established by Kongsberg Maritime.The test is based on a method called "rotating disc method", where the materials are exposed to aggressive cavitation for long periods.The cavitation specimens are weighted at given time intervals to determine the material.Here, the DED material from the PQ was compared to cast and heat treated duplex stainless steel and bronze material currently used in various components.

Figure 3.
Testing method and specimen geometry for BPQ and PQ.Specimens were extracted from three directions X, Y and Z, where X and Y are in the building plane and Z is the building direction perpendicular to the deposited layer.

Effects of DED process and HT on microstructure and material properties
In Table 3, Figure 4 and Figure 5, the measurements of BPQ and PQ specimens after heat treatments are compared to the measurement of BPQ specimen without HT.As can be observed, the HT contributes to fulfilling the requirements on tensile elongation and Charpy impact energy.While all specimens answer the requirement on the ferrite content, the BPQ and PQ specimens after HT show lower deviations in the measurements compared to BPQ without heat treatment.This can be explained by the microstructure homogenization due to the heat treatment process.The effects of the heat treatment on the microstructure are illustrated in Figure 6.By comparing the microstructures of the heat treated BPQ specimen with the BPQ specimen without HT, it can be observed that the austenite phase in the heat-treated specimen has coarser microstructure compared to its morphology in the as-built specimen.In the as-built specimen the austenite phase within the ferrite grains consists of sharp needles and non-homogeneous distribution of ferrite [7], while after HT, the austenite phase is more homogeneously distributed within the ferrite matrix.This contributes to improving the resistance to fracture as demonstrated by the increased values of the tensile elongation and Charpy impact energy.At the same time, the coarse microstructure of the austenite phase in the after HT might explain the reduction in tensile strength and bending force due to their effect on the deformation mechanism (see Figure 4 and Figure 6).This is consistent with scientific studies on dual phase steels showing that the refined microstructure results in increased strength [11] [12].The variation in ferrite fraction between the different specimens might also have an effect, however, it is difficult to conclude on such effects based on the current measurements.
The microstructures in the heat-treated BPQ and PQ specimens are quite similar (see Figure 6).The material properties measured for the BPQ and PQ specimens after HT display small deviations as demonstrated in Figure 4 and Table 3.This indicates consistent DED and heat treatment processes for the investigated geometry.Figure 4.d shows the effect of the actual section transition in PQ specimen on the heating and cooling rates.During heating and cooling, the temperature measured inside the thick section of the PQ specimen has a small deviation compared to the temperature measured inside the thin section.This was not found to have any effect on the resulting microstructure which was checked in both positions.The BPQ specimen as built by DED shows lower performance in the building direction, while the heat-treated specimens demonstrate small deviations between the material properties measured in the different directions (see Figure 5 and Figure 6).The more isotropic behavior of the heat treated material is explained by the homogenization of the microstructure resulting from the HT process.The only exception is the impact energy, which has more deviation in the testing results, but the deviations for the heat-treated specimens are much lower than the deviation for the as-built specimen.

Heat treated DED material versus heat treated cast material
Tensile, bending and cavitation erosion tests were also performed on duplex stainless steel specimens manufactured by casting followed by standard heat treatment.The testing measurements of the heattreated cast specimen are compared to the testing measurements of the heat-treated PQ specimen in Figure 7.The heat treated PQ specimen demonstrates slightly higher yield and tensile strength and improved resistance to cavitation erosion compared to the heat treated cast specimen.At the same time, the heat-treated cast specimen has considerable higher tensile elongation.This might be explained by the refined microstructure obtained by DED compared to the coarse microstructure obtained by casting.The required bending force seems to be less affected by the differences in strength and elongation, only a slightly lower bending force is measured for the heat treated cast materials.More detailed observations of fractured and eroded surfaces of specimens after tensile, cavitation and bending tests are planned to characterize the effect of the microstructure on damage mechanisms for stainless steel manufactured by DED process versus as cast material.

Conclusion
This paper demonstrates the great potential of DED manufacturing process combined with optimized heat treatment on duplex stainless steel to achieve the material properties and component performance required for critical maritime components, such as waterjet impeller.An optimized DED process and HT enabled achieving the required microstructural balance, tensile strength, tensile elongation, Charpy impact energy and hardness.The DED with HT process improves the resistance to cavitation erosion compared to heat treated cast specimen.This is expected to have a positive impact on the component performance in working conditions.

Figure 1 .
Figure 1.Simplified illustration of the manufacturing workflow showing the relations to the requirements and the qualification testing.

Figure 2 .
Figure 2. Evaluation steps, showing the manufacturing specimen produced for qualification testing at each step.

Figure 4 .
Figure 4. Tensile curves of a) BPQ as built compared to after heat treatment, b) heat treated BPQ specimen compared to heat treated PQ specimen.

Figure 5 .Figure 6 .
Figure 5. Bending force -displacement curves of a) BPQ as built compared to after heat treatment.b) heat treated BPQ specimen compared to heat treated PQ specimen.

Figure 7 .
Figure 7. Heat treated DED material compared to heat treated cast material in term tensile stressstrain curve (a), bending force-displacement curve (b), and relative weight loss normalized by the maximum relative weight loss on bronze as a reference material (c).

Table 1 .
Minimum requirements on mechanical properties of the material

Table 2 .
Process parameter optimized for DED manufacturing of duplex stainless steel

Table 3 .
Ferrite fraction, impact energy and hardness measurements of BPQ specimen as built by DED, heat-treated BPQ specimen and heat-treated PQ specimen.