Skip to main content

High Productivity Electron Beam Melting Additive Manufacturing Development for the Part Production Systems Market

Final Report Summary - FASTEBM (High Productivity Electron Beam Melting Additive Manufacturing Development for the Part Production Systems Market)

Executive Summary:
The FastEBM project is aimed at developing the Electron Beam Melting (EBM) process, a form of 3D printing in metal, so that it can be deployed for airframe components. 3D printing is attractive as it efficiently uses material compared to machining parts from solid, and is at its most attractive for high cost materials; FastEBM focused on titanium alloy as this material is expensive, difficult to obtain and machining costs are high compared with aluminium alloys for example.

The EBM process is slow which coupled with the capital cost of the equipment makes this production route expensive. Most of the production cost is due to equipment depreciation, so the higher the productivity, the lower the cost per part. For machined parts, the ratio of material bought to that in the final component, termed the buy-to-fly ratio, is used to categorise components. The highest buy-to-fly (BTF) ratio components (e.g. BTF = 20) are the most expensive and the EBM process can produce these parts at a similar cost. However, the majority of components have a BTF of less than 15, and to address this market the production rate of EBM must be increased so that the part cost is lower.

Speeding up the productivity of EBM machines required modifications to the equipment – so that a higher power beam could be generated. In order to investigate higher speed processing the EBM process was simulated in 3D using a Lattice Boltzmann method that has proven to be effective in 2D.

The equipment modification required new electron gun design capable of producing up to 10kW beam power, but with a resolution (i.e. focused beam spot size) similar to the current system with a maximum power of 3kW. The gun was designed using a finite element modelling software package. The cathode material was selected for its long life and high intensity of electron emission at low temperatures. The materials used were lanthanum hexaboride and cerium hexaboride. The gun was built and tested at TWI and then transferred to Arcam for process trials.

Modelling of the process was developed by the University of Erlangen-Nuremberg and the simulations were run on the SuperMUC supercomputer. Process variables were explored for high speed melting and many simulations were carried out, investigating parameters at up to 10kW beam power and results indicating whether the EBM build suffered from porosity, or layers swelling from being at elevated temperature, or whether the build was sound. This work has shown that a process window is open to allow higher productivity, and that process window has been graphed to allow it to be implemented in the new processing machine.

A process cost model has been developed to allow production costs to be calculated for a variety of components.

The work has been disseminated at several international conferences and in journal publications. The work has been presented to EADS who have provided technical advice for the target application.

Project Context and Objectives:
The EBM process is conducted in a powder bed in a vacuum box. A powder feed system deposits even layers of powder of some 100 micron depth. An electron gun is then used to heat the powder and then melt a layer of the component being built. The beam is moved across the powder bed by magnetic deflection coils. The powder bed moves down and a new layer of powder is distributed and then the next layer is formed.

The EBM process allows bespoke parts to be made with a very low wastage of material. The FastEBM project is focused on the aerospace industry and in particular titanium airframe components. At present these parts have high material costs due to the quantity of material machined away. Although EBM could build these parts with little material wastage, it is at present unable to address the larger components, and is too slow to be economically viable for this market.

The FastEBM project aimed to develop a higher power electron beam without compromising the resolution through keeping the beam diameter small. This would allow the beam to build more rapidly, and many beam spots could be used by time multiplexing the beam over different parts of the powder bed. Low aberration deflection coils have been developed to allow the beam to be deflected over a larger area to build larger parts. To understand the beam-powder interaction a numerical-physical model of the process has been developed. A key objective of the project was to increase the productivity of the process by a factor of 5 through combining the model with the new hardware.

There have been studies carried out into understanding and modelling the interaction of powder beds with laser beams. Lasers interact with just the first few atomic layers of material. Electron beams interact with matter quite differently; the electrons penetrate into material and for smaller particle sizes may pass through them, partially losing energy causing heating of the powder particles. There has been some work on the interaction of electron beams with powder beds, and the research work in FastEBM built upon and complemented the work carried out on laser interaction with powders.

Some three years prior to the project, the group of Dr. C. Körner at the University of Erlangen-Nuremberg had developed a 2D numerical model based on a Lattice-Boltzmann approach to model the beam powder interaction for the EBM process. This model represented the starting point for the planned modelling and simulation part of the project. Beam-powder interaction modelling aimed to increase understanding of the way the process works and provided a means to EBM build large components more rapidly.

Technical barriers:

The project addressed significant technical barriers preventing rapid EBM:

• This development will enable more efficient use of the time multiplexing of the beam across many working positions that will allow multiple melt spot operation to build the component concurrently. This requires higher electron beam power, whilst it is necessary to maintain the focused beam diameter to ensure that build resolution is not compromised. Normally higher power beams have larger diameter, and the component accuracy would be poorer.
• Large components require the beam to be deflected over greater angles in order to cover the extremities of the powder bed. Current deflection systems cause beam aberration that compromises build quality for components any larger than 200mm diameter. Low aberration deflection systems will be required to build larger components.

• EBM with multiple beam spot operation and higher beam power required further understanding of how the beam heats and melts the powder and mitigation of potential thermal runaway, where the build quality would be compromised due to thermal distortion or excessive melting due to the high thermal input of rapid build processing. The project developed a two and three dimensional physical and numerical model of the powder bed and beam-powder interaction, which was used to define the optimum beam qualities.

The project aimed to develop new technology in the design and fabrication of a new form of EB gun and deflection system, and the creation of knowledge to predict and control its use in production of components.

The development and knowledge created and protected in this project aimed to allow the SME collaborators to exploit the advantage gained through the development of vastly superior EBM production systems.

Project objectives:

The objectives of the FastEBM project were:

1. To design build and demonstrate a new electron gun design capable of producing a 0.2mm diameter focussed spot at beam powers up to 10kW (i.e. greater than 3 times the present power capability) by milestone M6 (month 17)
2. To design a low aberration deflection system that enables parts of up to 400mm diameter to be built without compromising the build resolution by milestone M3 (month 6)
3. To develop a model to simulate the beam-powder interaction at beam powers up to 10kW and use this to generate beam parameter and powder quality requirements for optimised high productivity processing by milestone M6 (month 17)
4. To integrate the high power gun and low aberration deflection system onto a prototype production system by M5 (month 13)
5. To produce a production model and feasibility study quantifying machine and processing costs by M7, month 16
6. To demonstrate build of a part using the new electron gun, deflection system and beam-powder simulation model at a production rate at least 5 times higher than presently possible by milestone M6 (month 17)
7. To develop understanding of the generation of low aberration deflection fields for high power electron beams to be sharply focussed through the use of 3D magnetostatic analysis by milestone 3 (month 6)
8. Develop novel beam-powder bed modelling and numerical simulation methods to allow derivation of rapid processing parameters and control techniques for many simultaneous beams using models of powder bed additive manufacturing with electron beams, by milestone 4 (month 5)


Project Results:
Following extensive modelling activities, a novel electron beam gun type has been selected as being suitable for the project application. The gun has been designed is currently being manufactured.

Deflection systems have been analysed and a new deflection coil has been designed to give higher angle deflection with lower distortion. An astigmator also has been designed (a weak octopole) that is integral with beam alignment optics immediately after the anode plate. The astigmator will work in conjunction with the deflection to optimise the beam intensity across the powder bed. Material for these high frequency elements has been selected – and identified in forms and sizes suitable for construction of the gun column elements. The electrical drives for the amplifier have been considered and a design proposed.

A CAD mechanical model has been developed for the complete gun column, to meet the dimensions and designs from the electron optical models.

The 2D models of the powder bed interaction with the electron beam have been extended to a 3D model of limited volume. In this task UEN implemented the numerical model including the beam definition, movement and absorption and the generation of the random powder bed. By the end of this task UEN were able to simulate single process simulations with a moving beam in a random and parameterisable powder bed.

The modelling and results were reviewed and agreed by the project partners. Results are encouraging and some comparison to experimental results has been carried out using high speed video of the process.

Further tests on the RF Plasma Gun identified that this technology, although offering very significant benefits to the EBM process, would not be ready to deliver a 10kW beam for this project within the schedule. Consequently, and in accordance with the risk register, a contingency plan was initiated to design a 10kW triode gun that could be readily integrated with the Arcam system. The 10kW electron gun was simulated, designed, built and tested at TWI on an experimental system. The gun was integrated with a 10kW high voltage power supply and a filament and bias power supply – all operating at 60kV. During the tests the gun peaking curve was plotted over a range of cathode heating. Also the beam diameter was measured at a working distance representative of that used in the EBM process. These tests showed that the gun was performing as predicted by the computer simulations, and that over most of the beam power range the beam generated was compliant with the project objective i.e. to produce a beam of full width at half maximum (FWHM) of less than 200 microns.

The gun and power supply were then transferred from TWI to Arcam, where they were integrated with an EBM system to allow the gun to be tested using Arcam’s standard procedures, albeit at beam powers of 10kW rather than 3kW. Tests on this system allowed gun peaking curves of up to 10kW to be plotted. Beam melt tracks were made and measured to give an indication of the beam width. The beam width was measured with a beam probe device mounted in the EBM chamber.

3D parts have been manufactured by the prototype FastEBM system to allow comparison with present equipment.


The 3D model of the process has been refined and has been used to investigate the best processing parameters particularly at high beam powers. The process window – defined by examining the process line energy vs the electron beam scan energy – has been explored in the simulation for the higher velocities that will be used at higher beam powers.

The modelling and results were reviewed and agreed by the project partners. Results have been compared to experimental results in the UEN Arcam EBM machine.

Costing of the process has been developed within a production model to allow the cost of parts made with the system to be estimated.

Potential Impact:
The project has developed a faster electron beam additive manufacturing machine that will be capable of making components up to 400mm long. Processing will be further developed with this equipment design to enable manufacturers to produce a range of bracket components that will be supplied to airframe producers.

The airframe market is dominated by Airbus and Boeing. The total market size for airframe brackets is from €1.1 billion, in 2014, growing to €1.3 billion in 2018. The segment of this market that can be addressed competitively with the EBM process is some 30% of this, as some airframe brackets are too large to be economically produced in this way, and many are manufactured from aluminium alloy, where the economic use of material has less effect on costs. The addressable segment of the market is therefore €305 million, in 2014, growing to €356 million in 2018.

Although this project has been focussed on airframe components, there are a number of secondary markets for similar components that are required in a wide variety of designs, mitigating against any downturn in the aircraft industry. These include engine and aero-engine parts, sports automotive, power generation, food processing and chemical plant.

Airframe brackets are currently manufactured through machining of forged pieces, or through machining from plate. The key disadvantage with this manufacturing route is that most of the metal purchased is machined away. The industry uses the phrase ‘buy-to-fly’ (BTF) ratio to measure the wastage of raw material that is not in the final product, and values of 15:1 by mass are not uncommon i.e. the bracket manufacturers buy 15 times more metal than the mass of products they supply.

Some 46% of the airframe bracket components supplied today have a BTF ratio of 10:1 or higher (see Table 1 in attachment). Although the metal machining debris is recycled, there is an energy cost, and a constraint in metal supply, especially for metals such as titanium. The aerospace industry purchases 40% of the global supply of titanium, and the price of the metal depends heavily on the level of aircraft productivity. The outputs of the project would allow significant material savings to be made, resulting in much lower costs for the manufacturer, more efficient (lower energy) production, and more efficient use of resources. In contrast, Fast EBM has a BTF ratio of 1:1.05.

A comparison of the price for FastEBM manufactured parts with parts machined from plate is given in Table 2 in the attachment.

A typical part mass of 3kg has been selected. The EBM processing cost has been calculated to include the cost of capital and operational costs.

The selling price for an EBM manufactured airframe bracket will be some 54% lower than the conventionally machined component. For example, as shown in Table 2, a bracket of 3kg mass, machined from solid, has a current typical price of €1,691, assuming the approximate median BTF ratio of 10. In contrast, the EBM product price will be €786, assuming a build rate of 1.5kg/hour – a target in FastEBM (see Figure 1 in attachment). This will be highly competitive, and will increase in competitiveness if the titanium price increases as is expected and has historically been the trend.

The key objective of the FastEBM project was to increase the build rate of components to allow parts to be produced at a competitive price. The sensitivity of the viable price to build rate is shown in the diagram above – and it can be seen that by bringing the build rate up from the present achievable rate (for large components) of 0.1kg/hour to above 0.5kg/hour for a 3kg component allows the addressable market share to increase from 0.1% to some 70%. At a build rate of 0.5kg/hr the EBM process is price competitive with machined components with a BTF ratio between 5 and 10. If the build rate can be increased to 1.5kg/hour the components produced by EBM will be competitive with machined product with a BTF ratio of 5 or higher – this represents 99.9% of the parts produced for the airframe building industry.

The cost savings for the product are made entirely in the manufacturing stage for the majority of products. The savings associated with reduced metal stock are not significantly offset by the price of titanium powder – this is some 3 times the cost of plate by mass. The powder cost is higher than plate due to the additional gas and energy costs in its production. Consequently, with a 15:1 buy-to-fly ratio, metal cost savings will be in the ratio of 5:1 for the EBM process cf machining. Most products produced by EBM are expected to be of the same form as machined product so there will be no further energy (ie aircraft fuel) benefits. However, EBM does give opportunity to fabricate hollow and optimised structures that would be expensive or impossible to machine. This could offer light weighting benefits in fuel consumption.

• At present some 10% of the aircraft weight is titanium, with 8% of the 10% in the airframe (the other 2% being in the engines)
• With increased use of composite in passenger aircraft there has been increased use of titanium due to their compatibility (similar expansion coefficient)
• As a rule-of-thumb, a 1% saving in aircraft weight equates to a 0.75% saving in fuel consumption
• Therefore fuel saving opportunities are of the order 0.1%, and over the life of the aircraft would accumulate

An estimate has been made of the sales achievable from the project outputs 5 years post-project. The market data assumes that a typical bracket price is €786 and that a typical aircraft fabrication requires 1,200 brackets, of which 360 are titanium and could be made by the process. Using the market data on new, greater than 100 seat aircraft demand in European (EU) and rest of the world (RoW) a total market size has been estimated in 2014 of €305 million growing to €356 million in 2018. We estimate, conservatively, that the outputs of the project could be used to address 1% of the European market in 2015, climbing to 8% by 2019. For export to the rest of the world, 1% would be captured in 2016, climbing to 6% in 2019.

Using these assumptions, we calculate that sales arising from the project range from €0.7million in 2015 to €23.1 million in 2019, with €16.1 million being exported out of Europe (see Table 3 in attachment).

Using the FastEBM process will result in greatly improved revenues for the SMEs. The process is knowledge-based and uses material efficiently. This will enable the consortium to compete with producers of machined components, increasingly sited outside of Europe. This fabrication route offers sustainable and profitable production.

Efficient use of rare and expensive materials has a number of advantages. The salient benefits of the FastEBM project are summarised in Table 4 in the attachment.
The project has resulted in the development of a prototype high power EBM machine, processing modelling and demonstration of its operation. Some further development is required after the project in order to qualify a wide range of parts, and to market the parts and processing.



List of Websites:
The project website (http://www.fastebm.eu) was established on 6/2/2012. The website has a public area – initially explaining the objectives and potential outcomes of the FastEBM project. It also has a member’s area – for members of the project consortium, which is accessed with a user specific password. Here there is
• A file repository for holding project documents
• A message board
Enquiries regarding the FastEBM project should be addressed to the coordinator, Ulric Lungblad at Arcam AB: ulric.ljungblad@arcam.com