Quality Conference (2015): Cost-Benefit Analysis – Automotive Industry

Quality Conference (2015): Cost-Benefit Analysis – Automotive Industry


Our next speaker, Dr. Ann-Christine Falck, a senior researcher at Chalmers University of Technology in Sweden.
She’s going to discuss Cost-Benefit Analysis from the automotive industry.
Well, hello everyone!
First I will tell a little about myself and my background because it of course has influence on the research that I’m doing.
Well, I’ve been working in the automotive industry for about 20 years.
First in production, in the paint shop, the assembly shop, the body shop, engine shop and so forth.
And I cooperated a lot with production technicians and people in production, and also carried through
a large amount of training for people in production.
And after that, I’ve been employed directly by manufacturing engineering at Volvo car corporation as an ergonomic specialist and a manager.
And as such I’ve been involved in all new car programs, I’m responsible for
all kinds of ergonomic requirement specifications, checklists, and standards and everything, and also for training of manufacturing engineers and project people through the years.
In 2009, I made my doctoral thesis, it was about as you can see here about assembly ability, product quality, productivity and health at work.
And that was at Chalmers University of Technology at Gothenburg.
So my focus is to look at ergonomics from a holistic view to sort of say.
I’ve been collecting my data mainly from the car industry, and we have tried to collect data in other places as well but it’s difficult because they don’t have
the available data that is need to make cost calculations the kind that I was going to do.
So first if we look at some statistics, this is from Sweden from the work environment authority, and this is what it looks today.
The red part of the circle is musculoskeletal disorder, and its not too much difference between men and women today,
it has been greater difference earlier, and it has decreased through the years, that means the number of cases has decreased.
But instead, the pink sector has increased, which means that there are more psycho-social disorders
today then it was earlier, and I think, probably there is a mixture of psycho-social disorder and musculoskeletal disorder, but they go for the main reason where they do the statistics.
As you can see as well, the other sectors also differ between men and women and that tells us that they don’t have the same kind of work.
Women are predominantly working in the health and care sector, while men are usually working in the industry and with technology matters.
Which you can see from the statistics.
I tried to get figure sorter from Europe, but it’s difficult because it’s so scattered, but MSD’s are also a major work related disorder in Europe as well.
As in all industrialized countries for that matter.
This is another way of looking at it,
when we see what happens, the problems start pretty early when people come to work
and they proceed through the years, and when we come up close to the retirement age, many people leave.
That’s up here, because they cannot work until 65 years of age, because they are to say worn out and people have too many problems, so people leave too early.
And the problem is then, how will we solve this since people in general, at least in the industrialized countries, are getting older and older,
and we cannot afford their pension, and we’re supposed to have younger people to do this kind of job but they are not.
So how do we solve this problem, it really is a critical problem if we look around us and see what it looks like.
MSD’s has a lot of consequences, I’m not going into the individual and society matters here
because it takes a bit too much time, but I will concentrate on what impact it has on the company.
And if people don’t feel well, they leave the company or they become ill
which results in rehabilitation costs, and in Sweden the employer is responsible for rehabilitation,
that means adapting the work place and the work to the worker. So it’s a lot of cost and efforts to do that.
You loose competence when skilled people leave.
It has effect on quality which I will share, and you also have loss on productivity if you have to redo a job again, that is repair things that have gone wrong before they can be sent out to the customers.
Below you can see some figures, they are a bit old now, but I think you can get some sort of an idea
of the cost. If you have a sheet metal worker, he cost almost a million Swedish Krons to replace.
You have a welder, it cost about half a million Krons to replace.
Spray painter almost similar.
And the cheapest as it looks when you look at this is an ordinary operator, instead there are many thousands of them, so when they leave it’s a big cost actually.
And these figures I got from the economics department of the factory.
So despite numerous ergonomics standards, regulations and actions through the years,
like EU standards, league of demands, company standards, ergonomics assessment tools, scientific facts and reports,
ergonomics compayence, that is the European agency for Safety and Health at Work that does that in Europe.
Poor ergonomics is still a major problem, why.
Is there a lack of research?
Well no, we have quite a lot of research, that should be sufficient.
Is there a lack of knowledge among decision makers and designers, well, there certainly is.
Too late actions? Yes!
Too late actions that don’t solve problems and are too expensive.
Corrective actions considered too costly? Yes, often decision makers think so.
And they want other solutions instead, like work organized rotation and facilities and that kind of thing.
Are ergonomists not involved in early development phases, well generally they are not, it’s quite rare actually.
Is there a lack of models for calculation of benefits in economic terms? Well, I would say there are for practical use and that is what I tried to create.
To find out why
these standards and all those ergonomics documents didn’t work,
we thought, well, maybe we should ask the people who are concerned about it.
So we interviewed 64 engineers in product development, in five different companies,
and they answered 43 instructed questions about ergonomics, quality, productivity and how it looked upon these things.
And the results were very clear.
Standards and legal demands are insufficient and they don’t change much, they help a bit but not enough.
Profitability calculations, that means hard facts are often required.
Preventive actions are difficult if ergonomics competence and cross functional cooperation is missing.
Sustainable assembly solutions require ergonomics knowledge, adapted requirement setting and working methods in design.
And assembly ergonomics design principles should be included in all technical training, we brought that up earlier but it could be further discussed how to do that.
And of course, ergonomics specialists should be involved in early development. That was the main conclusion of this study, there were a number of others as well.
So if we look at this, we have seen from a number of studies that poor product design
are the cause of 60-70% of all ergonomics problems.
And the process work stations design and work organization, correspond to 30-40% of all problems.
And this is a number of consequences and what has been discussed through the years a lot, are the health effects which we have mentioned today as well.
And of course that costs when people are becoming ill and leave the company or need rehabilitation or have to be replaced.
It also causes quality loses, and we have heard a little about it already.
And it causes repair and scrap.
And if you have to repair and do things once again you have productivity loses.
And all these things,
increase the total cost of the final product, and more the ergonomics statuses.
So instead of looking at the individual operator, which we have been doing for a long time, it doesn’t lead us any further, you have to have a systems approach to these things.
So it is about human technology interaction, so if you have insufficient interplay between humans and technology,
you are likely to have musculoskeletal problems and quality impact, resulting in productivity impact and resulting in reduced profitability.
Our studies and other studies have shown that there are common causes of ergonomic problems and quality issues, so if you have
products that are tough to assemble, you’re likely to have errors or thing going wrong.
If you don’t see what you’re doing, the same thing.
Poor accessibility result in both things.
Combined force and precision demands, the same again.
Too high force and weight demands, it’s quite well known.
Time pressure is also well known.
Then we have something called operator dependent assembly.
That means that, if a job is supposed to be correctly made, the operator has to have special skill. And if he has not got that, things will easily go wrong.
High repeativity, strenuous work postures, and so forth, that’s the cause of both ergonomic an quality issues.
This is an example, I have numerous examples but I couldn’t bring all of them here.
But this is a manufacturing engineer with whom I cooperated a lot.
And we were supposed to try to assemble, test to assemble a rod antenna as it was called, as it shows here.
So the design had thought that we could
take that rod antenna with that platform and three cables, and pull those through that tiny little hole in the roof.
And the working posture looked like this, and it was amazing that they believed that, and it was completely impossible so we said, don’t bring that in.
And they insisted,we said don’t bring it in, it will cause immense problems for both humans and quality.
So we brought all these people down to the special place in the plant
and they could test for themselves.
And they discovered of course that it was completely impossible to have that.
So big job started with people involved from paint shop, body shop, engine shop and so forth, and a lot of people with tooling so we tried one solution after another.
And finally this was the solution that was decided.
It was a semi-automatic assembly machine that put it on, and then we use a shock fin antenna solution instead.
And that solved all the problems and we got one cable left inside that an operator could attach when he was in the car doing some other things later on.
So from being a really disaster, this went to be a very good solution.
And in fact, another researcher copied this solution because they had the same problem in their design departments.
Well because you have to prove sometimes that it’s good for, well it’s good to make a change for a poor solution, you have to change it.
So then, I tried to develop a model calculation for the cost of poor assembly ergonomics and you can find that article in the International Journal of Industrial Ergonomics.
It was presented there in 2014.
The objective of this study was to show the relationships between assembly ergonomics, quality and action costs, and develop a model to be used by ergonomists and engineers.
So the method was to log manual assembly errors and scrap,
in the company,
both in the plant during assembly and on the market after the car had left the factory.
So for that purpose, 47 manual assembly task were corresponding to about 47 000 cars,
and we divided it into high, moderate and low load level tasks .
And for 38 of these, we could make cross calculations because we couldn’t use all the data to do that.
And to obtain all the data,
I had to go into a number of quality systems, starting with manufacturing engineering, I’m going through these in detail.
But I started there and then went through production, they have a system called attack, answer to all car questions it says, you could find very useful data in there.
And I also had a lot of cooperation with a material coordinators who provided me with scrap material.
And then the cars that went out of the market were followed by quality engineers there, who reported errors from all over the world that were collected, those errors that had to do with assembly were chose.
And so forth, and I didn’t do the job myself, I had some help as well, but a lot of collecting of data.
And the results show that, the errors at high ergonomics load level
actually were almost 8 times increased compared to the error tasks at low ergonomics load level.
But what is surprising is that also the moderate load level showed an increase of errors.
Which means that probably,
these solutions had been accepted and one believed that they were good enough, but obviously not.
Another thing that was very striking and we were very surprised about was that, 87% of all of these errors were geometry related.
That means, they didn’t fit, they were not connected, they sat in the wrong place and so forth.
And these geometry simulations earlier on had been considered as good from design, but when they came down to production, obviously they were not went handled by the operators.
And these geometry errors resulted in 89% of all action costs, and that’s quite a lot.
I’m cooperating with a PhD student who’s working with assembly geometry and it has given some interesting results actually.
So when we look at the action level of these errors, well the best ways to fix them as soon as possible.
That is online by the operator who did it, or maybe some other but on the same line, that was set as one, that was the best option.
If they didn’t discover or could not fix the error, it was a bit over 9 times
more expensive to fix then offline duty, disassembly and troubleshooting and things like that.
And when we looked at the errors that had slipped through all the control systems and were found in the market, it was even 12 times more expensive.
So that means that, the cost decreases very quickly, they later define the errors because a lot of disassembly, change of parts and so forth happens.
And the action time online then was on average two minutes, and offline about ten times as long that could vary a lot.
So the,
model then, I can go to the next, uh sorry
calculation model looks like this, but I think I go through it
like this instead. So the errors fixed online
times the average action time per error, there were a number of different errors,
was calculated, and the errors fixed offline was calculated with respect to the action times there, on average.
We had the cost for scrap items and parts and components, I got that from the material coordinators.
Some cars couldn’t be fixed immediately but had to be transported an put up in the yard, so the transportation time and the handling time was also calculated.
The quality system, well they had owed it, that means they picked out every week a number of costs and went through them in detail.
So those cars that had errors that
were related to our assembly task in the study were collected as well.
As errors belonging to series of cars that were stopped from being sent out of the factory.
Then additional efforts, that is extra staffing, meeting and control, etc, that could be needed but I couldn’t find that for this study though.
And then we have the cost for work related injury and sick leave rehabilitation as well, that was where it all started.
We also calculated cost market, it was quiet easy because you get a bill from external repair shops from all over the world, so it’s just all put together.
And then you have cost for recall and repairs are serious areas like bring your car to the repair shop immediately because there is something wrong with the brakes.
That is very very expensive, but luckily enough, I didn’t find any errors.
And the you have of course the cost for lost brand image, bad image which I haven’t got the figures for but of course there are some costs if the customer chose to well chose another brand the next time.
This was not included in the study but should be calculated, we did that during many years and there was an economist at Stockholm university
Paula Liukkonen, I don’t know if you’ve heard of her.
She helped us with this some years ago and it’s been used ever since.
Here I have the problem, it was pleaded cover insides the coupe.
It’s rubber, very big rubber mat that was supposed to be assembled, very heavy and difficult to handle.
It was supposed that they wanted to bring that in, and then they asked us to do a health cost calculation.
Then we looked at how many operators will be affected by this solution in our two plants then.
And about 30 operators would be affected.
And then we had to calculate the risk of some people getting hurt because some people manage everything and other people get problems, you never know for sure.
So the risk was considered to be about 40%.
Then you have to exposure time today, was different in both factories.
And then we estimated based on earlier statistics and so forth, the number of operators that could be concerned and how long they were expected to leave and the cost for that.
Then some people would probably leave because they couldn’t manage the job, we had the cost for that.
And finally the most troublesome part is the people who are still there but are not being able to do a full time job because they have to go for treatment and they are
sometimes doing some other jobs because they can’t do the ordinary jobs and so forth.
So they are not 100% efficient and there was a cost for that as well.
So in this case, we could have a total cost and split up that cost on the number of cars that were supposed to be assembled,
300,000, then you’ve got to do the cost per car, which is a bit over 5 Swedish Krowns which was considered very much.
Everything above a Krown was considered much, so of course they changed the solution.
So the conclusion of this study was that the high load level increased the errors, risk of errors almost 8 times, but also the moderate load level increased the risk quite a lot.
And there is a very strong relationship between physical load level and assembly quality.
The later errors are found, the most costly they are to fix.
And ergonomic actions can in fact be supported by holistic profitability calculations.
And there is then a great potential for increasing both product quality, productivity and decrease number of MSD’s, at the same time with earlier elimination of ergonomic risks.
That’s the first study.
So this is just to show that, the earlier you go in, the cheaper it is, and in early project phases it’s most optimal, but I’ll show you a better picture after this so you understand a bit better.
This is taken from the car industry, when I was working I was involved over here between 3 and 4 years before production starts.
And ergonomic requirements were hand shaked with the design engineers, they had to fulfill them and when it was not possible we discussed how to solve it then.
Very early I had a big possibility of influencing change of solutions, and major changes can be made because nothing is built, everything is done in virtual reality and simulations are made with computer mannequins now.
But half time in the project, important decisions are taken so the influence to change is decreasing.
When you come to production everything is built up and it’s very costly to make changes if possible at all.
This is just a checklist for load test ergonomics and the manufacturing engineers and the project people
had to fulfill everyone of the factors here. If it was not possible,
well the first thing was to change the product, the last thing to change was to send it down to production, sometimes it was a combination of both.
But this actually worked and this checklist was developed in cooperation with manufacturing engineering some years ago.
Now I have an example from Ford Motor company, I think it’s so old so I dare to show it now.
I think we, they complain, well, they have made a lot of improvements I would say because
it’s a big difference from then, but when they started looking at this
they found that they had a huge cost every year due to ergonomics, it’s really huge cost.
And that is what I call direct cost that is built form the insurance companies for treatment and rehabilitation of people.
But when they looked at indirect costs, hidden costs, like quality of use, replacement of workers, production losses, tooling changes, double ups and so forth,
they found that the cost was at least three times as high.
But the problem is that the majority of indirect costs are hidden, it’s not so easy to find and companies don’t usually log this and then it’s difficult to say anything.
So another study was
well also related with possibly before the first study actually and it presented in a scientific paper in 2010.
Then we studied about 24000 cars and also according to high, medium and low physical load level.
And again, there was a really strong relationship between poor ergonomic assembly concepts and the quality risks was increased,
but what was very interesting to see was that risk concepts,in many cases were already known before starter production but they brought them in any way.
And then about 80% of all the tasks at high, medium and low physical load level showed quality risks and they brought them in believing that production could solve them but they could not, not satisfactory at least.
So here we have some similar figures again, it was a bit earlier then so the risk was a bit higher.
I must say that I have worked a lot with these matters and have improved situations quite a lot.
But one interesting thing here is that, the action cost in the market
was in my study here almost 11 times higher then in the assembly shop and that’s exactly that figure that
some design departments say have come to as well, so obviously these figures were correct.
One sever problem with all of these is that there must be adequate registration system to find all of this data, and,
in many companies there are no data to do this kind of study and that’s why we have been able to do the study in just one or two companies.
Awh we can skip that one. So, uh
The continuous improvement cycle for the ergonomics is, when a project starts up, you should see what problems do we have today in production.
As sort of a feedback, and bring them in and try to solve them at the same time because there is a very good option in early phases.
And at the same time, identify the risks in the new project and then consider those,
the product, the process, the workplace, the individuals working there and the work organization, all ready then.
You don’t know everything but you should start to consider that because you can do that afterwards.
Then of course, starting with risk evaluation and ergonomic simulation in different physical tests in cross functional teams.
This was a nit picky process of course, testing and redoing it again and so forth.
And when one thought it was good enough, it was brought into production and
then we followed up the results and things that didn’t work well was brought back again.
So when you have done that a few times, you have got rid of many problems.
I would also say, what is the time? Do I have time left? Yeah!
There is another study that is not published yet, but, which I think is very interesting,
because a lot of these engineers talked about complex assembly tasks. What is that I said? Well, they couldn’t say.
But it was in my back head all the time so, this,
64 people in those 5 companies were asked to define assembly complexity, what is that.
And I got a number of answers and I put those together.
And the background was that because of the poor quality, I don’t know if you know this, equals to 10-40% of an ordinary companies total turnover.
A huge amount of that turnover is quality costs.
So I thought that I could use this complexity criteria to see if it had impact on assembly areas and action costs as well because ergonomics is not everything.
And these criteria are here.
If there are many different ways of doing the task, things are likely to go wrong, that’s what the interviewers answered.
If there are many individual details and part operations to keep track of, if there are time demanding operations, if there are no clear mounting positions of parts and components,
if there are product accessibility or hidden operations.
And here is ergonomics, if there are poor ergonomic conditions implying risk of harmful impacts.
If there are operator dependent operations that require special knowledge to be properly done.
If there is a need that operations must be done in a certain order, you must connect that thing, before you do that thing before you do that.
If it’s like that.
If there is visual inspection of fitting and tolerances, that means subjective assessment of the quality results by the operator.
Accuracy precision demanding task, need of adjustment.
If the geometric environment varies that means that, the component is the same but the place, the position of the component could vary between the different products, different cars.
Need of clear work instructions, soft and flexible material, and lack of immediate feedback. A click sound or a compliance of reference points.
These sixteen high complexity criteria were used in a study.
And I don’t have time enough to go through that, but it will be published I hope during the spring.
And it showed that,
there was as significant correlation between quality impact,
operator dependent operations was really something that you should not have.
The requirement or the factor operators must be done in a certain order of significance.
Visual and subjective assessment was visible.
Need of adjustment was also very important.
Lack of immediate feedback and geometric variation.
But you don’t find ergonomics here do you.
No we were all surprised about that, why didn’t we find ergonomics in this list.
Well,
no, the problem was solved.
As well as we didn’t find hidden operations and we didn’t find clearance operations because they’re focused a lot on that.
So this is what this study shows, but this doesn’t mean that it could skip some of theses matters here because it could depend on the situation in the company.
But, we were going to go further and deeper down into this to see what happens when we go on in further studies with this.
But the interesting thing with this was that, these were experienced manufacturing engineers and design engineers actually listed this criteria and it worked.
They were very experienced, on average they had nine years of work experience in their field.
Ahh, that is all I wanted to say.
So we do have a few minutes for questions, if you want to, please step forward and us the mic.
I would start with a clarification on your last point. Can you put your last slide up?
Because assembly time came up again.
First a comment, like to me, all of those issues relate to the interaction between the operator and the system, and as such I would call them human factors.
So to me that is ergonomics as opposed to the strictly the injury risk posture problems right.
But when you talk about assembly time, is that the fraction of time from a one minute cycle, let’s say the duty cycle.
Yes, the average time was I think 77 seconds here.
But when it was longer then that it was causing more problems, so that’s not so difficult to understand because the longer the time you have the more errors you can make to say.
Right.
So you think it’s a question of the volume of the work as opposed to the operators trying to rush to keep up with the task time.
Probably yes!
We don’t know really, but that is what the statistic analysis showed, but could be a question mark on that one.
Questions from the audience?
Cause I’ll just go on all day here if you want.
You did a fantastic job of drilling out the data from these big car companies.
And clearly that was difficult.
What do we, you know, the challenge we seem to be having is that when we start looking at the smaller companies, less organized then the auto sector, which is probably out front in manufacturing technology.
How do we deal with those companies, because they don’t have that data available?
No, I have tried myself to do research in smaller companies but it’s difficult, well you could do interview studies, that’s okay.
But, I don’t really know, but what we have found in our latest study where I interview decision makers at very high level companies,
they don’t really
read research reports and they don’t go to conferences, and they don’t know about this.
So I think it’s a knowledge problem.
So with the engineers and managers you were working with,
did you find that their attitudes changes as they became more familiar with this issue?
Oh yes! Oh yes! It was like turning the hand around.
So I would say that training is the key to success.
Training and ergonomics and better knowledge.
We did that in the company years ago and it was a tremendous difference, we didn’t have to persuade and try to convince people any longer.
It just worked by itself.
Once they saw the linkage to quality.
Exactly yes.
So and that was much more successful then you’re earlier efforts with a safety based ergonomics approach.
Yes it was!
Because people started to think themselves and the engineers started to fix the errors themselves,
without asking, well, the had to consult us but they did a lot of work without needing our assistance.
Anybody want to weigh in?
Yeah Margo! Thanks You!
So I guess I would just echo that, but a lot of times our training is just focused on musculoskeletal disorder prevention, but engineers don’t know what to do with that.
What’s the criteria around that? What do I do I do in my design to use that?
So until you start providing design criteria, it’s hard for them to do anything.
It’s not that they don’t see that there’s problem with the musculoskeletal disorders and so on, but, translating that in something actionable would be difficult to do, that would my experience.
And I think we see nano-electronic sector, and it sounds like you’re seeing some of the same thing.
How then do you take the,
human factors elements and translate them in a way the design engineers can act on?
How we did that? Well I have, sorry I din’t think that I bring that along with me, but we had about 20 requirements covering everything.
And that was for instance, working distance should not exceed a certain amount of centimetres regardless of the length of the operators.
The weight demand was that it shouldn’t be over 12 kilos, ten times an hour close to the body.
But of course the design engineer couldn’t understand all of that, they cooperated a lot with the manufacturing engineers to interpret this.
So it was sort of a cross functional thing there, and we were involved in some cases when it was difficult.
So just to make sure we’re understanding correctly, when you say design engineers that’s product design?
Yes! Exactly, yes.
So cause I wanted to ask more, how do you get to those product designers? We find that when we go into the shop floor we talk to operations,
very very difficult to get to the more distant product design people who are maybe in a different city, different facility.
How did that play out for you in terms of
Well usually manufacturing engineers were belonging to a core engineering department at that time, they did the job but, it’s no problem, just lift the phone an call them.
I’m just joking, it was like that and then we decided meeting to talk, it was seldom a problem.
So in the Volvo case, the design engineers are in the same facility they’rte available.
Yes they are. That’s one thing that makes thing much easier of course.
Ok, and thank you. Margo I totally agree you know. How is it that we change our criteria from a posture relative to elbow height. What does that mean to someone designing a car or someone designing a process industry.
I don’t think we necessarily have good answers for that yet.
And that’s part of the gap that we see between what we’re doing in ergonomics and focus on postures and muscle fatigue,
and what the engineers are up against in their daily battles to get the product out the door and onto the market.
Yes, but we did simulations with computer mannequins and also we could show what their solution meant and that helped a lot of course.
And then we did a legitimate calculation of risks or costs or whatever was needed.
Often was enough.
Interesting, thank you very much. We appreciate it.

Leave a Reply

Your email address will not be published. Required fields are marked *