The future of metrology in manufacturing
Thursday 30 May 2013
Last November, the metrology community came together to discuss the future of metrology in manufacturing at the LVMC event, held in Chester, UK. The panel discussion, chaired by Trevor Toman of Coventry University, consisted of leading global manufacturing companies and representatives from academia, the National Physical Laboratory, AMRC and media, all of whom have a strategic interest in outlining the contribution metrology can make to manufacturing in the future.
The LVMC conference also included metrology equipment suppliers and delegates from a good number of UK and international manufacturing companies, including metrologists and manufacturing engineers who use measurement technology in their job. The panel’s objective was to create a roadmap – or least make a start in this direction, following the old Chinese saying, “To travel a thousand leagues, you must take a first step.” The fundamental assumption underpinning this exploration is that the future of metrology in manufacturing is all about meeting industrial needs. It’s about metrology as a business tool to answer questions that users have on a daily basis. What follows is a discussion and dialogue between the panel members and the conference delegates.
Floor: What we would like to do is to express the value of metrology to senior management, that metrology is not just a cost: that there is a lot of knowledge to be gained from doing measurement.
Alan Olifent, Jaguar LandRover, (AO): I’ve found it very difficult to explain to people why I’m investing so much money in that measurement facility. At JLR that are adding up to some large values. My answer to that is it only takes one failure to blow that kind of money. A case study, for example, was that of the Toyota accelerator pedal which happened a couple of years ago, The cost to that industry, just in rework terms, was something like $120 million. The cost of failure of measurement, failure of process, failure of control, to Toyota was over $4 billion. Now it’s very hard to quantify cost of avoidance as compared to cost of saving because you don’t plan for failure such as the Toyota accelerator pedal, etc. But that’s the difference you see. You have to say, I’m confident in the engineering systems, I’m confident in the validation, what the engineers have designed – that’s where you get the value for metrology from.
Floor: If we are to avoid Toyota type failures, shouldn’t the metrology spend be recognised by the shareholders of a company in their annual statements?
Tallion Edwards, Boeing (TE): When you state that metrology as a separate item to shareholders, that it’s not part of manufacturing, that’s a problem in itself. Sometimes, this has been self inflicted by quality departments, adopting an attitude that says: “We’re here to find your problem” and that’s clearly not the case. Metrology reduces the risk to the program or, at least, makes you understand the risk. So, whether it’s a recall risk, or some insurance risk, the metrology value is to understand what that risk is and to make a business case based on those results
One of the big problems is that we stand alone. We stand out there as something you spend money on as opposed to something that’s integral, so that the design, manufacturing and metrology/quality teams realise that they are all part of the same manufacturing process. Quality should be integrated into manufacturing process.
Reg Marsden, Rolls-Royce (RM): We’ve talked about risk here. What we also need to talk about is how we move forward. If you can’t actually measure something, you don’t know anything about it. That’s one of the fundamental things. So if you want to improve your manufacturing, and to minimise the amount of inspection or checks you need to do, you need to have confidence in those processes. So really, metrology is about getting the processes right which you’ve also got to produce the product at the end.
Floor: Metrology isn’t just a check on a process. In many factories now, it’s integral to the process and in terms of value, it often helps to use metrology to set things, to fabricate more accurately and eliminate tooling. To say measurement is a check or a balance on a process, is giving it short shrift. There are a great number of measurement systems out there that aren’t measuring as a check but are integral to the process of fabricating something. Regarding value, if you are eliminating a lot of things that you used to have to design, for example in the Boeing world they’ve eliminated a large amount of tooling by using metrology instead and (reduced) the cost of engineering, that’s a huge cost saving.
AO: The issue is not one of metrology; it’s an issue of management. The best analogy is in football or soccer terms: a lot of the management I’ve had to deal with over the years have regarded the quality department’s metrology as the goal keeper. In reality, the department is actually the defence because if it gets through to the goal keeper, you’ve failed – your processes have failed, your manufacturing systems have failed. You use your quality systems to make sure it doesn’t get to the point where you have to take special measurement action or anything else.
For me, metrology is to find a way of eliminating tooling, or to do an in-process check on every product we make . In automotive terms, cycle times are 50 seconds or less on some products. So its every 50 seconds, you have got to complete your checks just to give yourself a little tick in a box, that’s metrology, not taking one sample a day out putting it on a CMM or other measurement system and checking it to the drawing.
It’s that education and awareness of what you can do with it is a bigger task than finding the latest whizz-bang or more accurate tool or better understanding uncertainty. The task is to get people to think about what they should be doing and pushing that forward.
As metrologist, as a function looking at quality, I want to be able to give the best answer but if they don’t understand the answer I give, then I’ve failed anyway.
TE: It’s not just an issue of management. It’s also an issue of ignorance. We know our business, the design teams know their business and the manufacturing teams know their business. Largely, we don’t know each others business. There was a good example of that earlier (at LVMC12) in the talk from Mark Jones from CERN. Talking about when business asks for an unreasonable number from the measurement process (perhaps because they don’t really understand the physics of the process). The measurement people have to go back to business and say that’s really expensive to implement. They have to have a conversation across two different vocabularies to try and decide where the best business decision sits between the best number for design and the best number for measurement. It’s ignorance on both our parts, not knowing the limitations that people have to deal with. And I’ll put it more on the team rather than on any individual, which is why education and (skills) certification in the industry are important.
Brendan Coyne- QMT magazine (BC): The value metrology delivers can perhaps be seen as residing upstream at design for manufacture of product and process stages within a model-based manufacturing and Lean quality management context. Metrology can provide the right measurement data for CAD/CAM designers for tolerancing purposes etc. In this way, it can be a value adding activity – knowing how to make something, and how well you can make it and on what machine and in which process gives you a database of manufacturing knowledge which you can store for the future.
Trevor Toman - Coventry University (TT): Design for manufacture which incorporates metrology and metrology tools into the original model - that’s where we need to go.
Floor: Design to manufacture. We are starting to use model based manufacturing more but this approach tends to tight, dimensional machining tolerances to the point that it almost assumes dynamic measurement capabilities as we move down the process. That’s not always available. As soon as you start working out uncertainties, it kind of throws the process internally. We tend to find that quite a difficult thing at times, trying to squeeze the best out of it that we can.
RM: Yes, this really is a problem for us (R-R). You often find in an organisation that a lot of investment has been made in the front-end CAD side and the machining. Then it comes down to, “Oh, yes. We do need to make sure that we are measuring this properly“. Unfortunately, this comes down at the end of the n’th stage. It’s back to hearts and minds, really. Getting recognition within the organisation of the importance of measurement, that’s really where metrology comes into it. And it’s the benefits you can get from it. What I really try and do in my role is to get designers to recognise what we are trying to do in metrology. i.e. to feed the information back to help the designers to make the products better. It’s closing the loop. It’s very important that we are all joined up.
Floor: Do the tolerances (generated by CAD) need to be that stringent? As industry is getting better machines, more precise machining, the question asked is: why can’t we measure to that size? But have we gone too far in this direction? Don’t we need to sit down with the designers right at the beginning of the process with the smart (digital ) model as ask what the right dimensions (and tolerances) are.
TT: If we are producing something to a design tolerances of +/- 1.5 microns, or whatever, for a number of years, then, as measurement systems improve, we suddenly discover that actually, we haven’t been producing these things to +/- 1.5 microns. We then have discussions to say now we have validated it, we can’t do it. The point being is that good design affords the largest tolerance to allow us to produce within certain risk elements. What we often see is that the measurement system is nowhere near as good as the production facility. The product is good and may have been out there functioning perfectly well despite our inability to accurately measure the product or process. So now, as measurement improves, we can say to the designers, here’s evidence that the tight tolerances aren’t required. If we loosened them, we could make the product a lot cheaper. In that respect, good measurement is a good business tool.
Fundamentally, we shouldn’t be designing things we can’t measure.
AO: The tolerances specified on a drawing cover all sorts of variation. That includes variation from your measurement process called measurement uncertainty. If your measurement system is the most variable part of the process and you are still getting good products – who cares?
Jim Heley - AMRC (JH): One of the issue, in terms of not understanding what we are producing, is that we think we are producing things to a tolerance but, in reality, we are not, and we subsequently to try and change the situation. This can be a big problem in industries, such as aerospace where you have qualification for processes. That can be a big limiting factor for taking metrology forward in automating processes and advanced processes. Once you start to understand that you are not producing to a certain tolerance, that raises a lot of problems and a lot of risks come out. That’s what we’ve found at the AMRC in Sheffield.
TE: We all know design engineers who say, “I wonder how accurately someone can machine that? If I can design it (tolerance it), can you manufacture it” But how many of them are asking the question, “How accurately can we measure this?” and that’s part of the process. You are using up part of the process capability and uncertainty. Nobody in the design community is asking that question right now. That’s a problem for us.
TT: If you don’t design anything you can’t measure, then at least we will be telling ourselves the truth. And we need to understand the capabilities of our measurement systems and tools, not just the quoted accuracy specifications.
TE: What should be in a roadmap for metrology? One of the things should be trying to find ways to educate everybody in the entire value stream, including management, design engineers, the metrology community and manufacturing engineers, on how to move metrology forward in the production process, bring it forward into the value stream. I don’t know how we are going to do that. What can we produce as a metrology community to make that message clear? That this activity has more value if the question is asked, how am I going to manufacture this and how measure before they put dimensions on parts? If those questions are asked upfront, then you have time to figure out what the best measurement tools are at the end. You are setting your tolerances based on capabilities, as opposed to the other way round.
JH: Do people actually buy metrology or do they buy solutions? We’ve talked about a machine tool and a metrology device as two separate things but can we merge things together? Should we form partnerships and create forums where people are working together to put in a production solution rather than a metrology device at the end of a production line? This is the future and we are moving away from that (model) already.
BC: Technology will have a tremendous impact on the future of metrology and its role in manufacturing. Just as technology is progressing, so too is metrology – newer technologies, such as non contact sensors, are leading to unprecedented tools, doing things unheard of years ago, from the macro level to submicron level and across an ever increasing range of surfaces and new materials. The key is to understand that they are technologies and not (yet) complete solutions.
Floor; An issue for SME’s is that there is a multiplicity of different metrology systems available that look very similar but achieve very different accuracies. What you can end up with is a metrology tool that looks as though it can provide the solution but, in reality, is not at all capable for the purpose for which it was bought. For large companies, this may not be a problem as they have the resources to evaluate the product and its application, using such things as gauge R& R etc. But for SME’s who don’t have these resources or knowledge, this can be a problem.
Ben Hughes - NPL (BH): That’s an interesting point. Part of the problem stems from the fact that often the specification of these metrology systems is documented in a very simple way. We tend to talk about an MPE (Maximum Permissable Error) of a coordinate measuring system. That tends to reflect a maximum permissable error in a length measurement. Very often, people aren’t using coordinate measuring systems to measure length, but something much more complicated, such as features and parameters. That in itself makes it very difficult to compare the suitability of the different kit that may look suitable for a particular application. Conversely, it doesn’t give a very good reflection of the capability of an instrument. For example, a laser tracker is, on the face of it, a relatively simple instrument but, depending on how you use it, you may achieve a factor of x10 better than MPE if you know what you are doing and can bias the sensors. There is certainly a case for more thought on how you can compare these systems, looking at task-specific system validation rather than just looking at MPEs, specification data sheets and listening to salesmen.. Apply more thought, more knowledge, more education
Floor: Should this education be driven back into the design side so that they know what metrology tools to apply?
BH: Yes and we also need to come up with better ways of specifying performance. The A + B/L formula that is typically used, though simple to understand, doesn’t reflect the coordinate metrology system’s performance in any real world situation. I think there’s a requirement for instrument manufacturers and standards bodies (not necessarily National Measurement Institutes) to look into ways of how these metrology systems are validated – how that validation is specified. But it’s complex: how to find a solution that apples to all coordinate metrology systems?
RM: In terms of getting the NPL and the resources to get people top understand, there is an opportunity for SME’s to talk to the NPL who have done a lot of work at looking at the different systems and can give some fundamental advice: there are lots of user guides. It does come back to education and people understanding the capabilities of the equipment. Things like the NPL certified Level 1 and Level 2 metrology courses are helping to improve the knowledge.
One of the key element s for the future of metrology is the education of the metrologists - not just metrologists, but everybody.
TT: What place does first principles metrology have in the future of the discipline?
RM: I think this is important. I like to look and see if people that are involved with measurement actually understand the fundamentals i.e. first principles of measurement. We have a number of people who come and write programmes on CMMs. I like to see is that CMM programmer has an understanding of first principles measurement. If you do that, (and that goes back to education) when you are looking at datums and where everything is coming from, you’ve got a proper understanding of the metrology.
JH: I think it might be going the other way. With the automation of everything, people are becoming less aware of metrology and how it works. I’m not necessarily saying that’s a bad thing, Look at maths tables and calculators, people don’t know their maths tables like they did 20 or 30 years ago, but they know they can just pick up a calculator and work out what 27 x 27 is or whatever. With the automation of metrology, with more the more complex systems that come out, people will have less and less awareness of how they work. And that’s a good thing. We should be looking at how to integrate these systems, and really almost making them vanish, but having the data there. The data coming out and we get that red light green light situation to produce a good product. I’m not sure which is the right way to go. We are always going to need experts – you need that fundamental metrology expertise, but is it going to be widespread or is it going to be less widespread?
TT: I guess what I was trying to get at was: it is very easy to press buttons on an advanced measurement system and read numbers to sub-micron levels, But if you don’t understand what’s behind that, are we in danger of losing our ability to assess the situation? So our metrology roadmap for the future of metrology in manufacturing would include: education, training, qualification of metrology users, closing the loop with designers, and understanding the application capabilities if different metrology equipment, looking at metrology as a holistic process, not just equipment and people in isolation, integrating metrology into the design and manufacturing process, and metrology as a solution. Key to any successful role for metrology in the future will be management’s perception of the value of metrology.
Finally, we haven’t got the all answers but then we didn’t claim we would have.