Evaluation of Additive Manufacturing Feasibility in the Energy Sector: A Case Study of a Gas-Insulated High-Voltage Switchgear

Featured Application

The application of this study is all aluminum parts in high-voltage devices and the feasibility of considering additive manufacturing as a manufacturing method.

Abstract

In recent years, additive manufacturing (AM) has made considerable progress and has spread in many industries. Despite the advantages of this technology including freedom of design, lead time reduction, material waste reduction, special tools manufacturing elimination, and sustainability, there are still a lot of challenges regarding finding the beneficial application. In this study, the feasibility of replacing traditional manufacturing methods with additive manufacturing in the energy sector is investigated, with a specific focus on gas-insulated high-voltage switchgear (GIS). All aluminum parts in one specific GIS product are analyzed and a decision flowchart is proposed. Using this flowchart, printability and the best AM technique are suggested with respect to part size, required surface roughness, requirements of electrical and mechanical properties, and additional post processes. Simple to medium complexity level of geometry, large size, high requirements for electrical and mechanical properties, threading and sealing, and lack of a standard for printed parts in the high voltage industry make AM a challenging manufacturing technology for this specific product. In total, implementing AM as a short series production method for GIS aluminum parts may not be sufficient because of the higher cost and more complex supply chain management, but it can be beneficial in R&D cases or prototyping scenarios where a limited number of parts are needed in a brief time limit.

1. Introduction

The conflict between fast-rising global energy demand and climate change is a major challenge that requires significant science and technology innovations. Advanced manufacturing has the potential to reduce greenhouse gas emissions and pollution significantly and shorten the time-to-market [1]. Additive manufacturing (AM) is defined by the ASTM society as the “process of joining materials to make parts from 3D model data, usually layer upon layer, as opposed to subtractive manufacturing and formative manufacturing methodologies” [2]. In general, AM eliminates the design and fabrication restrictions of conventional manufacturing (CM) methods to a large extent. As an emerging and transformative technology, AM and 3D printing offer great potential for the energy sector by decreasing the lead time, improving energy efficiency, eliminating the cost of molds and special tools, and allowing for increased customization of parts [3]. According to estimations, the global power system of 2050 will require four times today’s generation capacity and will need to transfer three times as much electrical energy [4]. Therefore, more sustainable manufacturing technologies (such as additive manufacturing) and advanced design concepts will play an important role in the minimization of material usage and reduction in waste material generated. Despite the positive aspects, several challenges eliminate the adaption of this technology to the industry application. Production volume, standards compliance, product quality and properties, post-processing, size limitations, and metal AM cost are some of the most important drawbacks in the path of using AM [5].

In the energy sector, AM has already demonstrated its potential in producing customized parts for wind turbines, gas turbines, solar energy production equipment, fuel cells, and batteries [6,7,8,9]. In addition, the possibility of employing AM technologies on-site to produce spare parts for the replacement of damaged components in oil and gas equipment and facilities has been considered. Although the reduction in production downtime and replacement costs are considered benefits, the future of AM in the oil and gas industry needs further research in the area of new materials development processing, the improved surface finish of AM fabricated parts, enhanced fabrication speed, and parametric optimization to improve the mechanical properties of the fabricated components [9]. AM in energy will undoubtedly bring the expanded application to nuclear energy. Numerous investigators will join early adopters in applying existing commercial AM methods and developing new approaches to nuclear fuel fabrication [10]. Moreover, AM and 3D printing in power electronics and converter applications have been explored, focusing mostly on inductors and copper-printed windings [11]. Printing of magnetic materials with respect to the topology optimization for electrical machines has been discussed [12]. The application of 3D printing for the manufacturing of high-voltage transformer insulation has been investigated, and several challenges have been identified when considering 3D printing. The presence of internal voids in the printed insulation is one of the main issues for high-voltage transformer insulation [13]. In another study related to AM for HV application, additive manufacturing of polymeric components used in an HV apparatus was discussed, indicating that because of printing limitations, such technology can be used only for non-critical parts [14].

Considering the previous publications, it is clear that AM for high-voltage applications has not been investigated deeply, especially for metallic alloy parts. This could be related to specific requirements in such applications, as well as the large sizes of the products. Therefore, this study focused on the feasibility of AM in high-voltage products and considered a gas-insulated high-voltage switchgear as a case study. AM markets, available AM technologies, and powders for aluminum alloy printing as the dominant material in GIS products were evaluated. A decision flowchart was proposed considering the printability of each aluminum component in GIS products. To verify the accuracy of the proposed flowchart, a unique product, GIS-ELK-04/145 kV was selected as a case study. The results show that the proposed decision chart is a useful tool for business unit managers to choose the best AM methods due to their requirements. However, from an economical perspective, AM is inefficient for GIS aluminum parts in production lines. This technology can be a good fit for spare parts manufacturing and prototyping because of the short lead time and elimination of the need to manufacture molds and special tools for low-volume orders.

2. Materials and Methods

This paper investigated how and if AM technologies can be a sufficient method in energy industry production lines, spare parts management, and prototyping. An initial investigation indicated that most components in GIS products are made of aluminum alloys. Therefore, this study focused on aluminum AM and 3D printing methods and their potential. As the first step, a careful evaluation of the AM market and available technologies for aluminum alloys was performed.

Table 1. Evaluation of the AM market and available technologies for aluminum printing.

Method	Alloy Type	Challenge/Limitations
Powder Bed Fusion (PBF)	Al-Si-Mg	Most mature AM methods for Al alloys. Challenges limited to the lack of standardization, high cost, and limited alloy type.
Electron beam additive manufacturing (EBAM)	All weldable aluminum, e.g., 4043	Not ideal for aluminum. Because of the vacuum chamber, there is a need to add other compositions with Al material, essentially creating a new alloy.
Ultrasonic additive manufacturing (UAM)	Aluminum 6061	More efficient for dissimilar metal combinations and embedding sensors or fiber optics into the part and/or have complex internal channels.
Laser hybrid welding (MIG, MAG, powder, etc.)	All weldable aluminum	Requires machining to obtain the final geometry.
Laser metal deposition (LMD)	All weldable aluminum	No available service provider for aluminum alloys.
Laser Engineered Net Shaping (LENZ)-DED	Al 4047	Requires machining to obtain the final geometry.
Binder jetting	Al-Si-Mg Al-Si 6061	Size limitation. At the R&D level.
Reactive additive manufacturing (RAM) technology	A6061-RAM2	Same as PBF but a different alloy, which results in faster printing and higher mechanical properties.
Material jetting (aluminum liquid metal AM)	4008 A356	At the R&D level.

shows the results of the market evaluation and the pros and cons of each technology. However, the results are based on the claims of the companies, without printing samples for all techniques. Evaluation outcomes show the maturity of PBF methods and the significant challenges of other AM technologies for aluminum alloys. In addition, Directed Energy Deposition (DED) and hybrid printing methods that combine near-net shape printing and machining to achieve the final design and tolerances can be an optimum solution in some cases. The selection of the proper method needs some consideration including the complexity of design, post-process requirements, and required surface roughness.

As the second step, a decision chart was proposed to consider the printability of each aluminum part without deep consideration of cost aspects (

Table 2. Decision chart of AM feasibility for aluminum parts in GIS products.

Parameters	Different Levels
Complexity	High	Medium	Low
Size (cm)	Up to 40 × 40 × 40	Up to 155 × 155 × 110	Above 155 × 155 × 110
Aluminum alloy type	Casting alloys AXXX	Machining alloys AXXXX
Quantity of order	1–2	Up to 10	Routine production
Standard requirements	Standard mechanical properties tests/micro-structural analysis is enough	Pilot installation required	Strict standard and regulation
Re-design possibilities	Yes—significant	Yes—minor changes
Lead time	Urgent	Normal	Not important
Average surface roughness (µm)	Above 6.3	1.5–6.3	Below 1.5
Additional machining/ post-processing	Without sealing area and/or thread	With sealing area and/or thread
Electrical conductivity (%IACS)	Below 25	25–45	45–60
Ultimate tensile strength (MPa)	Up to 220	Above 220

Green areas indicate the sufficient printability criteria and red areas show disincentive parameters of using AM.

). As a case study, one product, a gas-insulated high-voltage switchgear (ELK-04/145 kV), was considered to verify the accuracy of the proposed charts. GIS-ELK-04/145 kV is the smallest GIS product in Hitachi Energy that is more AM-friendly in comparison with other models. However, the results of using the chart/flowchart on the candidate case study can be generalized to the other GIS products because of their similar designs, functionality, and technical requirements. The key parameters to determine printability include the level of complexity, size, type of aluminum alloy, quantity of the order, standard requirements, re-design possibilities, lead time limitations, required average surface roughness, additional requirements for machining or post-processing, and electrical/mechanical property requirements.

The complexity classification is based on unique figures that AM enabled us to design. For example, TPMS/gyroid structures, internal surfaces, and any complexity that cannot be covered by CM are categorized as high complexity. Simple complexity includes any pipes, tubes, simple cylinders, plates, boxes, etc. Any geometry that can be manufactured by CM methods and needs some manufacturing steps is categorized in medium- complexity level.

According to the market evaluation and decision chart, a decision flowchart was proposed to be implemented in the case study (Figure 1). In this step, the presented flowchart was used for all aluminum parts of GIS-ELK-04/145 kV to obtain the fraction of parts that can be printed. Next, printable aluminum parts were investigated in a cost analysis considering routine production, spare parts production, and prototyping for R&D, and technology development projects.

According to the proposed flowchart, if the part complexity is medium and the size is smaller than 155 cm *155 cm *110 cm, two options can be considered. One option is the Directed Energy Deposition (DED) method with consideration of any necessary heat treatment and the selection of alloy type due to the required electrical/mechanical properties. The second option is printing by PBF-based methods and considering any potential heat treatment, machining, or post-processing. The example in

Table 3. An example of a path through the flowchart.

Geometry Complexity	Size Limitations and Post-Processing Requirements	Final Decision
		Two options: DED methods like WAAM. (There are no special requirements for electromechanical behavior, so there is no need for HT or special alloys.) Hybrid solution: Printing and then machining for threading and surface roughness requirements (below 6.3 microns).

shows the path through the flowchart for one part, and the red arrows show the direction of selection due to the technical specifications and requirements. Other examples are summarized in

Table 4. Some of the evaluated aluminum parts in GIS-ELK-04/145 kV.

Part	Weight (kg)	Challenge	Decision Based on the Flowchart
	0.767	Multiple threads. Surface roughness below 6.3 microns.	DED. Hybrid solutions (PBF + machining).
	0.554	High electrical conductivity. Surface roughness below 6.3 microns.	PBF with Al-Si-Mg + HT + PP. DED with aluminum 5XXX + HT.
	0.370	Internal gear. Mechanical properties and fine surface roughness of printed gears.	DED with aluminum 5XXX. Hybrid solutions (PBF + machining).
	0.04	Multiple threads	DED. Hybrid solutions (PBF + machining).
	0.427	High mechanical/electrical properties requirements. Including thread.	Hybrid solutions (PBF + HT + machining). DED with aluminum 5XXX + HT.
	63.9	Large part manufactured by casting in a steel mold. Multiple threads. Fine surface roughness.	DED. Hybrid solutions (PBF + machining).
	11.6	Large part manufactured by casting in a steel mold. Precise long through hole.	PBF

.

During this study, more than 1500 parts were evaluated one by one. The following important facts need to be considered for this evaluation:

• The range of size, average surface roughness after printing, post-processing, and mechanical/electrical properties, which are considered for categorizing the parts, are subject to change over time, and the presented data are based on the date of writing this paper;
• Current aluminum parts produced by sheet metal manufacturing methods are not considered in this investigation;
• The binder jetting method is also under development for printing aluminum, and it is assumed to be a cost-efficient method because of the significantly high feed rate. However, the technical evaluation of this technology is under investigation in a separate study. In addition, the currently available building size is small, and it does not cover the majority of the GIS aluminum parts;
• The proposed flowchart considers printability, so cost-efficiency is another item that is discussed in

  Table 5. Cost analysis of AM vs. CM for candidate parts.

  Number	Part	Weight (kg)	Current Cost with CM Methods (USD)	Estimated Cost with AM Methods (USD)
  1		0.767	14.6	620 (without threading)
  2		0.554	70	450 (without HT)
  3		0.370	85	309 (without machining of gears)
  4		0.04	6.1	36.9
  5		63.9	Total cost: 647 - Cost of mold design and production: 11,350. - Steel mold will be used for 2000–3000 parts. - Cost of casting + machining per part: 641.	~ 40 k–50 k
  6		11.6	Total cost: 136 - Cost of mold design and production: 8352. - Steel mold will be used for 10,000 parts. - Cost of casting + machining per part: 135.	~9 k–10 k

  separately. However, some logic of being beneficial with AM is thought out in Table 4 as well;
• The level of complexity is defined manually and case by case. However, there is a potential to use deep learning to classify the complexity of geometry. Here, a convolutional neural network (CNN) is suggested for this application.

After the feasibility study, it was clear that AM technologies are available for more than 90% of GIS aluminum parts. Although printability is not a problem, one of the most important parameters that limits the replacement of conventional manufacturing (CM) methods by AM is cost. Therefore, as the next step, some of the parts were selected for cost evaluation. The presented cost evaluation has the following considerations:

• The cost of AM is estimated by the weight of the part, and it is considered to be USD 0.8 per gram for the SLM method and AlSi10Mg alloy;
• The cost per part with AM is considered as outsourcing to the service provider and without any special contract for large-volume production. The current cost of CM methods is also based on outsourcing to an external supplier but under a large volume production contract;
• There is no re-design consideration, and it assumes the same design for conventional and AM methods. However, the possibility of re-designing for most of the parts was not significant.

3. Results

3.1. Printability of Aluminum Parts in GIS Products

• According to the proposed decision charts (Table 2), the green area indicates the parameters with a good match to AM as a production method. During this study, and with an evaluation of more than 1500 aluminum parts, it is observed that there are multiple challenges through the path of productization with AM. First, significant numbers of aluminum parts are large and bulky without the remarkable possibility of re-design because of the surrounding elements. Consolidation of multiple parts was also limited because of the multiple material printing requirements. Second, there are costly and time-consuming tests that need to be performed for printed parts before installing them in customer sites except for some aluminum parts used as joints and connections. Another challenge of implementing AM in this industry is the additional required processes such as threading or post-processing for sealing areas. The additional required processes factor is crucial as it will increase the supply chain management risks and cost of the final part. Finally, the required electrical and mechanical properties are significantly high and need additional heat treatment or special powder. Table 4 shows some of the evaluated parts and their limitations. It is clear that most parts are in the range of simple to medium complexity; however, because of their maturity and availability, PBF methods are one of the options for printing.

The evaluated parts shown in Table 4 are some examples of the analyzed parts and indicate the available AM methods and solutions for most aluminum parts in GIS products. However, more than 90% of the components need additional processes after printing, threading, surface roughness improvements, or heat treatment to increase electrical conductivity or mechanical properties. Therefore, supply chain management will be more complicated, and the benefits of AM resulting from one-step production with a short lead time will be ignored. In addition to printability, the testing and validation of all required standards is an essential step. The high-voltage power industry faces restricted standards and certification for the majority of parts. Pilot testing and going through the certification process is a challenging and costly process, so AM currently can be used for parts with less strict standards.

3.2. Cost Analysis of AM vs. CM for GIS Aluminum Components

The evaluation of the parts’ printability with the help of the proposed decision flowchart shows that there are available solutions to replace CM methods with AM techniques. However, cost can be a bottleneck for implementing AM in production lines. The final cost of printed parts can be compared with traditionally manufactured parts in two different categories. First, parts that are in the production line and considering AM as the replacement of series production. Second, spare parts and prototyping in technology centers and research-based projects.

3.2.1. Cost Analysis for Series Production

Cost analysis for short-series production needs to consider multiple parameters to compare the cost of printed and current products accurately. This evaluation should take into account all costs including storage, transportation, necessary tools for production, etc. Since the comparison in this study is based on the outsourcing of both production methods, warehousing and transportation can be considered as the constant parameters. The cost of the printed parts is estimated by the weight of the part, and it is considered USD 0.8 per gram for the selective laser melting (SLM) method, one of the PBF techniques, and the type of powder is AlSi10Mg alloy. This estimation is based on received quotations from multiple AM service providers. The final cost of the current parts manufactured by CM methods is also due to the outsourcing to Hitachi Energy manufacturing suppliers. Table 5 shows a comparison of the costs of the candidate parts with the AM and CM methods.

The results indicate that the price-determining parameter for AM is weight, and on the other hand, the cost of machining is defined by geometry complexity. Therefore, in the case of short series production with AM, the best candidates are the smallest ones with the highest complexity in design. Considering the parts in GIS products, small and complex aluminum parts that can be printed without any additional post-processing are not available.

Current CM methods are different because of the size, weight, geometry, and material requirements. Table 5 shows different parts with different current manufacturing methods to cover all aspects of cost. Part numbers 5 and 6 are manufactured by the casting process and machining. Although the price of special mold design and production will be added to the final cost, for series production, this cost will be divided into 2000 to 10,000 parts, and the share of the cost per part will be negligible.

3.2.2. Cost Analysis of Spare Parts and Prototyping

Spare parts management comes with a long list of challenges; high storage and logistics costs, overproduction caused by minimum purchase volumes, and long waiting times are standard and prevalent and, sometimes, the supplier no longer offers the part by the time it is needed. AM offers a solution here. Spare parts made from metals can be produced on demand by a 3D printer without tools and in the exact quantity required, without a minimum purchase volume. This avoids unnecessary overproduction and guarantees maximum customer satisfaction. The underlying idea is instead of resource-intensive warehouse logistics, the CAD data of the spare parts are saved and only sent when necessary. Moreover, as the pioneering technology leader, Hitachi Energy collaborates with customers and partners to enable a sustainable energy future—for today’s generations and those to come [15]. Therefore, besides the advantages of spare parts-on-demand, AM’s aim at clean and sustainable technologies is especially considerable.

Prototyping for technology and R&D centers also has the same situation as spare parts manufacturing in terms of lead time. The design and manufacturing of special tools and a long lead time of up to 6 months for each part in the development phase are significant challenges for designers. AM is a suitable solution that can be considered for prototyping with short order to delivery time and without any need for additional tools.

In this study, since the exact effect of the delivery time of spare parts to the customer is uncertain, we tried to make this comparison based on only the cost of the required tools and consumables. As a result, if only the cost of production is considered, AM and 3D printing are efficient for parts currently produced by the casting method and require the design and manufacturing of special steel molds. The reason is that in the production of spare parts and prototyping, the tool’s price is considered for one part, and this tool may not be used anymore.

According to the evaluations carried out in the case study (GIS-ELK-04/145 kV), between USD 8000 and 14000 should be paid for the design and production of the casting mold. If we assume the cost of the mold is USD 10,000 per unit on average and consider that the cost of AM and 3D printing is USD 0.8 per gram for aluminum alloys, then AM for spare parts or prototypes weighing up to 12–13 kg will save money. This savings in cost is in addition to the fact that the waiting time is reduced from 3–6 months to two weeks.

4. Discussion

In this study, a decision flow chart is proposed for the market evaluation of aluminum AM methods. To verify the accuracy of this flowchart as well as consider the printability of parts in high-voltage products, a case study is selected. GIS-ELK-04/145 kV is considered as a case study, and the proposed flowchart is applied to every aluminum component of this product. The results show that there are significant restrictions and challenges due to the replacement of CM methods with AM technologies. These challenges include high requirements of electrical and mechanical properties, threading and sealing areas in the majority of parts, and large and heavy parts without the possibility of significant re-design and weight optimization. Additional post-processing and heat treatment will reduce the benefit of AM methods because of the increase in supply chain management risks. Moreover, most aluminum parts in GIS products for small to medium-sized parts are made of aluminum 6061 and 6063 with high mechanical properties and proper electrical conductivity. However, the available AM methods are focused on Al-Si-Mg alloys (high Si percentage) with lower mechanical properties and lower electrical conductivity because of powder with laser-induced. In addition, the long process of standardization for some unique parts is ignored.

The proper testing and qualification of products is one of the most important tasks when applying additive manufacturing in the high-voltage industry. As mentioned earlier, the high electrical conductivity of metals is a must in most HV applications; therefore, in the selection of proper 3D printing technology, such material properties must be verified. In addition, the quality of the printed microstructure must be high without internal cracks and voids, minimizing the potential failure of the part during normal operation. Such robust and reliable function of the product must be assured for even more than 20 years. This means that long-term aging tests of the printed parts (e.g., creep behavior) should be performed. Currently, HV engineering has no standard procedures indicating how the printed components should be qualified. The most reliable procedure in Hitachi Energy is standard tests (type tests) for specific products. Since these tests and related certifications are costly and time-consuming, in the future, the development of standardization procedures will be required. Some sectors, including aerospace and automotive, have already established various consortia aiming at the development of such standards. Here, ASTM organization plays an essential role in establishing its Additive Manufacturing Center of Excellence (AM CoE) and, together with academics, industry, and government, is carrying out research and development to accelerate the development and adoption of 3D printing technology [16]. Such agreements are an essential part of the adoption of AM for energy sectors.

Another challenge in considering AM as a manufacturing method is cost. A cost comparison between AM and CM methods for some candidate parts shows the significantly higher cost of AM methods for series production. However, AM can bring huge benefits for spare parts management and prototyping. Short lead time, order-on-demand, and cost savings are the benefits of using AM for service units and technology centers. AM methods can be a suitable case to prevent engineers from manufacturing an expensive mold for one spare part or prototype. Besides lead time, even from the cost point of view, if an aluminum part with the current manufacturing method of casting + machining is less than 12–13 kg, it is beneficial to print the spare part or prototype instead of manufacturing the mold.

The presented study focused mainly on the application of 3D printing for manufacturing aluminum parts/ spare parts in GIS products; thus, for most of the components, the same design was used as that in traditional manufacturing. This approach is not cost-effective; therefore, as a next step, more effort will be placed on the application of topology optimization and design for additive manufacturing. Although design optimization for spare parts is impossible in some cases, it can be an efficient tool for short-series production. In this case, we can be efficient and more sustainable because of the reduction in weight and printing time and, finally, reduced cost. With this consideration, we can advance closer to the commercial application of AM and short-series production in the energy sector.

5. Conclusions

In summary, this study shows the following:

-

The proposed decision chart is a suitable tool for identifying the proper AM method for aluminum parts in GIS products.

-

There are multiple challenges in printing aluminum parts of GIS products that reduce the benefit of replacing CM with AM methods. These challenges include the size and weight of parts with no complex geometry, threading, and sealing areas in many parts, and high requirements of electrical and mechanical properties as well as surface roughness.

-

Considering the mentioned limitations, AM is not a reasonable technique for the replacement of CM methods for series production. Cost and SCM risk increments are the most important parameters that prevent this replacement.

-

Spare parts manufacturing and prototyping in service and technology center units are the best use case of AM because of the shorter lead time, no need to manufacture the mold, and limited order quantity.