Rapid Manufacturing May Deliver Carbon & Cost Savings
Consider this: The light-weighting of parts (due to honeycomb interiors) through additive layer manufacturing could be a much more important advantage of rapid manufacturing (RM) than the lack of tooling required. This is one of the early findings of a ground-breaking UK project.
Started about a year ago, and scheduled to run through 2011, the Atkins Project is a US$4.1M university/industry collaboration that seeks to revolutionize
(The project is primarily looking at metallic RM, specifically MTT Technologies' Selective Laser Melting (SLM) technology. Despite the relative immaturity of metallic machines relative to polymeric systems, metallic RM has a couple advantages. One is the ability to reuse excess material from metallic builds; in the plastic RM evaluated by Atkins, much of the excess material is lost, due to thermal material property changes. Metal RM parts also have mechanical properties that better match traditionally-manufactured metal components; plastic RM parts are still for the most part too different from traditional engineering polymers.)
In addition to industrial partners like Boeing, Caterpillar, and Bentley Motors, small UK-based rapid manufacturing consultancy Econolyst Ltd has a prominent role in the study. Econolyst is responsible for developing the software and processes necessary to control a "Grid-RM" supply chain. The company is also developing an online carbon and economic impact assessment tool to compare metallic RM processes against CNC machining & casting. (Also check out their side project: Per-Snickety, an enterprise dedicated to developing systems to 3D-print video game characters.)
Following is a recent email interview RapidToday conducted with Econolyst Managing Director Dr. Phil Reeves.
RapidToday: Where does the Atkins Project get its name from and what is the project's overall budget and funding sources?
Reeves: The name Atkins was a joke that stuck. Basically, the UK government was putting funding into Low Carbon Manufacturing research. When we first started talking about the project three years ago we referred to it as ‘low-carb' rapid manufacturing. The objective of the project is also about making light-weight parts. Because of the US low-carb Atkins diet, we jokingly started referring to the projects as the Atkins Project. After a few months, no one could find a better name and it stuck.
The project overall budget is £2.9M [US$4.1M]. This is made up from £1.45M from the UK government Technology Strategy Board and £1.45M from the industrial partners. The project is what we call a 100% cost project, in that all the partners have to spend £1.45M in total on the project to lever down the £1.45M from government.
Of the £1.45M of funding, a large percentage will be going to Loughborough University, where the low carbon production system is being developed. Econolyst accounts for approximately 10% of the industrial contribution.
RapidToday: What are the principles and drivers behind Atkins and why is additive manufacturing being investigated in this way?
Reeves: We believe that additive manufacturing could enable five fundamental economic and environmental business benefits, particularly for transport-related components. First, as we know, additive processes allow us to design parts differently. One of these differences is the ability to design with optimum strength-to-weight ratios. This has two effects: (1) the part requires less raw material and (2), the final part weighs less. This is important, particularly in aerospace, as weight equals fuel usage, cost and carbon. Also, materials such as titanium are expensive, so it makes no sense to purchase a billet of materials and then send 95% back into the supply chain as chippings. Also the design freedoms of additive manufacturing allow us to design and produce more efficient products, particularly those with fluid or gas channels. The two other benefits that we are investigating are the efficiencies of metallic additive process compared to casting or machining and also the configuration of the supply chain used to make parts, as we are no longer constrained by fixed assets such as tooling. This element of the project we are calling digitally distributed manufacturing.
RapidToday: Generally, what are the data exchange processes required for digitally distributed manufacturing? Is it as easy as, for all vendors that are signed up, checking first for geographic relevancy, then excess capacity?
Reeves: In terms of signing up vendors, it is far more complicated that just looking at geography and capacity. The analogy I give it that of CNC machining. There are literally millions of 5-Axis CNC milling machines globally, but ownership does not make you a validated supplier to Boeing for instance. But Boeing buys CNC machined parts on a global basis. What we are working on within the partnership is supply chain validation protocols for RM.
We are very much in the early stages of this at the moment, but we envisage a world similar to the current machining supply chain, where some companies will specialize in certain sectors (aero, auto, medical) and others will offer generic capacity.
In distributed RM the objective is to have multiple distributed suppliers that can utilize under capacity to keep costs down, but also where manufacture can be as close as possible to part consumption. What we are developing is an intelligent print queue to manage this process. The software looks at who are validated suppliers and their location, and then manages the flow of information within the supply chain.
Irrespective of manufacturing location, the overall objective of Atkins is to ensure that the life cycle carbon footprint of parts is lower than traditional manufacturing and that the life cycle economics of using additive processes are also equal or lower than traditional processes.
RapidToday: Has anyone tried to calculate the carbon and economic impact of building new distributed manufacturing facilities and rapid manufacturing equipment. Existing infrastructure may not be very efficient, but it is existing.
Reeves: Great question! To our knowledge this is the first time that anyone has attempted to calculate the carbon benefits of distributed manufacture and the life cycle impact of parts made using additive processes. There are lots of studies about ‘food miles' and purchasing goods that are grown closer to the point of consumption. We are using some of this thinking as our reference materials. However, when it comes to calculating the carbon footprint of let's say a CNC-machined aerospace bracket against a selective laser melted part over a 30-year life cycle, we are very much treading new ground. This is where the Atkins partnership is vital, as through the partnership we have access to a lot of traditional manufacturing processes, complex existing global supply chains and products such as commercial vehicles, motorsport components, prestige cars and long haul and short haul airliners.
In terms of the technologies themselves, efficiency is a vital consideration for Atkins as a project. MTT Technologies - which makes the SLM process sold by 3D Systems in the USA - is one of our partners. At present the best-in-class metallic additive processes use fibre lasers with around 20% efficiency. Hence, 80% of the energy going into the laser is lost as heat. As part of Atkins we are trying to understand where the losses in the manufacturing process are and how they could be re-used.
At present we are developing a web-enabled comparator for metallic RM parts which we will be beta testing with the Atkins partners this summer. This will allow them to make real time environmental and economic comparisons between RM and their current manufacturing processes. We think the key to RM adoption is in considering the environmental and economic benefits of the part over its entire lifecycle, not just the fact that RM mitigates machining or tooling.
RapidToday: How will the online assessment tool work? What major assumptions are made in the algorithms?
Reeves: The on-line assessment tool is based on the new British standard for carbon foot-printing and looks at the five main contributors to carbon. These are raw material, manufacturing production, product distribution, in-life product use and end-of-life product disposal.
In terms of raw material, we look at comparing the raw material used for conventional CNC machining (billet) or casting and compare this to the RM material (atomized Ti or Al). We know the energy used to produce both billet and powder so we have a starting point. We then look at the waste streams in the manufacturing process, such as swarf generated by machining compares to support structures generated by Additive Layer Manufacturing (ALM). Of course, metallic ALM has a fantastic material efficiency compared to CNC machining.
Within the software we also look at cycle times, power consumption, water utilization, mitigation of hazardous manufacturing processes and post processing. We then look at how the part moves within the supply chain between traditional manufacturers and the possible new distributed RM supply chain. We then look at how the part is used. If it is an aero part, how much weight can be saved using RM and how will this effect fuel consumption over the part life. Finally, we look at reuse and recycling.
One area we are looking at with RM is parts consolidation and the reduction of different material types, as this will aid end-of-life recyclability. In essence the software walks you through these five steps for both the traditional part and the RM part. It then provides both a carbon and economic comparison over the product life. This is important as even higher cost RM parts may have a lower life cycle cost implication to the customer if they can reduce in-life servicing costs or fuel usage by being lighter weight.
In terms of usage, we have a number of revenue models that we are looking at; for commercial reasons, I can't go into it in too much depth. The project partners will, of course, have full enterprise access. We anticipate for all other users it will be a pay-to-use software tool, most probably on a monthly licensing model. The aim is to get people running multiple scenarios through the system to look at different life cycle economic and carbon costs.
Our plan at present is to get a very-early-stage demo on-line for demonstration at the 4th International RM Conference at Loughborough UK in July 2009. However, the algorithms are based on a number of assumptions that we can only validate through experimental methodologies. These experiments are currently taking place within the partners, but we are hopeful that we will have their results within the next 4 to 5 months, so we can integrate these.
I can't really go into the assumptions, as this would in-effect describe the algorithm. Suffice to say, at present no one really knows what the carbon footprint of a CNC machined titanium aerospace part is, or for that matter an aluminum casting. However, we will later this year.
RapidToday: What are the educational implications of changing the way that we make parts and the training and retraining of thousands of engineers to use new RM design techniques?
Reeves: In terms of education, we are still very much in the early days of re-writing the design for manufacture rule book. The reality is that all additive processes have design constraints. Things like powder and support removal are the most simple to educate people about, then we have issues such as thermal mass, anisotropy, distortion, laser beam parallax error and so on.
At the moment, most of this knowledge is a ‘black-art' that is tied up in the heads of systems users such as bureaus. Some are cascading this knowledge out to their customers. We do run courses where we touch on design for RM (see www.rm-masterclass.com). However, it is limited. I know of a couple of books and a knowledge-based engineering software system under discussion, but we are a few years away yet.
RapidToday: Distributed manufacturing seems great for certain products, like video game characters, where there is presumably already distributed excess production capacity and no further post-build operations. I have trouble seeing it in automotive and aerospace, however, (except for replacement parts) because auto and aero parts still need to be assembled into a whole, and it seems inefficient to have hundreds of little assembly operations all around the world.
Reeves: Yes true - Grid RM works best where the production and consumption are within the same geographic part of the grid (if we are trying to be just carbon efficient). This would work very well for consumer goods manufacture and spare parts. However, please remember the cells of the grid can be geographically quite close and in some cases the transportation of the part may have only a very small carbon footprint during the entire life cycle of the part.
The grid also works if you have a production demand for, let's say, 1,000 parts and 10 validated suppliers each with capacity to supply 100. Some may be local, some regional, some national and some even international. What we are trying to do with Atkins is calculate the impacts (both environmental and economic) of using different suppliers at different locations, as this is where we believe there is a real economy of scale
Let me give you an example. Over its life cycle an aircraft component could travel up to 100-million miles. The difference between manufacture in Mumbai or manufacture in the jobbing shop next door could be 6,000 miles, that's 0.00006% of the life cycle carbon footprint of the part and that assumes the part was air freighted to its point of assembly in the first place. For such a part, location of manufacture is almost a carbon irrelevance. Hence, it becomes purely a quality, cost and delivery-based decision.
RapidToday: Also, regarding weight reduction, are safety issues a factor? With aero, the lighter the better, but is that true with auto too? Individual parts with latticed interiors may be just as strong as their solid counterparts, but is a lighter vehicle inherently less safe?
Reeves: On the contrary, we are seeing light weight lattice structures being used to also absorb energy, as they can be designed to crush in more controlled ways than solid or machined parts. Of course safety is a significant issue within the project, hence most of the parts that we will be trialing are non-safety-critical parts, with minimal loading.