FMEA Practical Tutorial

welcome to complexity made simple my name is paul allen and before we get into today's video well i'm just going to remind you the three books that are on sale drink tea and read the paper if you're a green belt in a black belt and you want simple instruction on how to apply your skill design of experiments for 21st century engineers and finally a statistical process control for small batch production they are all available from lulu.com and the links are in the video below welcome to complexity made simple my name is paul allen and the subject of today's newsletter where we're going to take a look at fmea failure mode and effects analysis and in particular we're going to take a look how to fill out an fmea a process fmea by the way so this is specifically a process fmea and we're going to look at how to fill out an fmea properly how to do it fast and how to do it to make money because that's what an fmea does it makes money and to be quite honest the fmea i'm you know i'm into six sigma and i teach people how to solve problems um a lot of the time i'm not using fmea to do that if the fmea the process fmea is used correctly at the beginning of the cycle six sigma would never exist six sigma is not needed if an fmea is done correctly so we're going to take a look at a process we're going to take a look at a process fmea so some people like to put a p in front of this thing pfmea the the reason why i picked this subject one of my clients is implementing a new manufacturing process so they are not going to use the traditional six sigma problem solving type techniques because they don't have a problem yet they are going to use fmea and they sent me they sent me this document so this this is telling me that there is a new guideline now for the aig fmea process this thing sort of tells you the benefits it doesn't tell you exactly how they're going to do it but the the benefits in this document one of the things they talk about is the fact that the new methodology and it says here it's a more structured approach it leverages lessons learned so it's building on control techniques that you've already got i guess but here's the one i'm really interested in it is error proofing driven it is error proofing driven now what that's basically saying is this is more about prevention than detection and that's really what a great fmea should drive you towards prevention rather than detection so let's just set the context up first of all fmead now an fmea in a real sense is basically a risk assessment so if you haven't done one before if you haven't done an fmea before and by the way most people hate fmeas they're normally slow sluggish soak up lots of time people just fill them in to keep customers happy and keep iso audiences happy but they never use them that a complete waste of time most fmeas think about your company is your fmea valuable that's what we want a valuable fmea so it's basically a risk assessment now what is risk made up of well it's made up of the severity of the error that you're likely to make it's made up of the likelihood and i think that's how you spell likelihood yeah likelihood not sure um the likelihood that it's going to happen and finally most fmeas talk about the detection controls how likely are you to spot the error and if you multiply the severity by the likelihood by the detection controls you work out what's the risk okay so if it was a health and safety risk assessment what's the risk that you're gonna have an accident you're gonna kill somebody that the the unit is going to fail and cause a severe problem what's the what's the risk now of course in a process fmea what's the risk well typically the risk is pretty straightforward the risk is defects take that one stage further if you let those defects out of your company the risk is dissatisfied customers and what will the satisfied customers cost you if you let them exist dissatisfied customers can cost you your business that's what the risk that's what the risk is now if you use an fmea properly if you use a process fmea properly and this is what i've advised my client to do he has a new process what i'm gonna get him to do is design the process you are going to design the process into the fmea form we are going to design the process straight into the fmea form and what's going to come out of the fmea is a control plan and then when you switch the process on bang it's right first turn that's the power of a process fmea you can make more money faster make good products faster you can go live defect free that's what that's what it's for it's it's the designer process um so let's take a look at the traditional fmea form and then we'll talk about the the importance of filling it out so here is the basic fmea form uh so on the left hand side look you can see we're going to design a process it says what's the process step all right so let me just show you a product that maybe we're going to use as a little example so let me let me just show you a product so here's the product that we're going to use as as a little fmea example so imagine you're setting up a production line to make a remote control for a television okay so you've always got to make the electronics you've got to mold the front you've got to mold the back you've got to mold the battery cover um you've maybe got to mold the rubber buttons then you've got to assemble the whole thing and then it it's got to work okay so that's the product that we are going to design the process for in actual fact i'm probably just going to talk about the battery cover just to keep this simple but you will see the procedure on the form so we're going to use this as a little example and then we can design it into the form okay so process step now then this has got to be in detail so something to avoid in the process step we don't just want to say something like mold battery cover all right so we could just put mold battery cover now there's not enough detail in that what you have to do is talk about okay we're going to make the battery cover so step one is maybe i don't know you're gonna maybe you're gonna weigh out material so maybe you're gonna weigh out how much plastic you want next step you're going to dry material feed to molding machine and so i'm going to blow it into the top of the molding machine feed to the molding machine then we're going to uh heat in the barrel then we're going to uh inject to tooling and we might talk about the cooling time eject from tool and then maybe pack packing boxes yeah so you can see the detail that's needed okay you can see the details nearly now if you look at that what am i doing well i'm effectively designing the process obviously for a molding process um it's it's a fairly generic thing so whatever molding activity i do those and how many lines have i got those eight lines could be standard every time i do this and this whole fmea could be standard every time i do a moulding activity but you can see there's some detail i've designed the process now if this was an assembly production line of course this would be more of a design function it might be more unique to making that remote control as opposed to actually assembling the tv which again would be a completely different fmea and you'd have to think uniquely you have a new product with a new design a new rear cover and maybe a new way of fixing the rear cover etc so the fmea would be more unique for a more unique product but for molding what we're looking at here could be very very generic okay let's go back to the form now the next box here this box here i'm going to color this thing in let's color this thing in red here because this is super important what is critical to quality what is critical to quality about this step so i'm just going to pick the drawing of the material so we could say it's the moisture content okay we could i would prefer to do it like this maybe and i'm gonna call it drying time so drying time is crucial what is critical to quality about this step drying time is crucial okay so what you've done is you put your thoughts down now you've designed the process you are saying which parameters are crucial to the quality of the product okay let's go on let's go to the next box now it says what's the failure mode now let me change something in the failure mode because this is how can the input fail so this is an input failure so in other words how could the drying time go wrong what could go wrong with the drying time what could go wrong with the item that's critical to quality what could go wrong with the input to the process because drying time is an input so if we go i'm going to say drain time is too short there is the failure mode then we say what's the failure effect now the failure effect let's put the output this is the output failure so you can see we've got two failure modes the failure mode how does the input go wrong the failure effect how does the output go wrong now if you get these mixed up this is where fmea start to go completely to pieces because if you put the wrong thing in the wrong box here so under the output failure let me just enter something in the output box a second we'll come back to this now so i'm going to get white marks in the molding and then maybe in brackets high scrap okay so okay so i've put i've got white molding i've got white white marks now in the plastic and they have a high scrap rate if i put that in the wrong box if i say the failure mode is that i'm going to get weight marks in the molding then i say what's the failure effect now you're having to think forward you're having to say well there's going to be a high scrap rate maybe we're going to be late with the customer because we haven't got enough parts maybe we're going to delay an order to a big customer are we going to lose a big order now you have to guess because you you put the failure effect in the wrong box you put the failure effect in the input box now now drying time is critical to quality what could go wrong with drying time the drying time is too short what effect will that have on the process a high defect rate due to white marks in the molding that is the proper way to do it inputs and outputs there is the six sigma principle loading clear in the fmea and by the way that was always there 20 years before six sigma ever existed fmei was floating about and inputs and outputs were always there people just never were never told how to do this but you have to get the input in the right box the output in the right box let's go back to the form now what we're going to do is we're going to give this failure effect a severity score now this is often from one to ten now the piece of advice here with the severity score is that you write your own table describing how painful this problem is so i'm going to show you some text from my book called drink tea and read the paper here is a severity table this is actually for products rather than for processes but i i use it because it's from the car industry but it tries to be nice and specific so they try to describe ranking 10 so look ranking 10. um when the potential fail amount affects the safety of the vehicle um you know and the problem could happen without warning so you're gonna get a problem without warning it could be a safety related problem so in other words you could injure or kill somebody and the failure will happen without warning that's a 10. if you go down to an outlook it says the vehicle or the item is inoperable with a loss a loss of primary function so they're trying to be nice and specific so that selecting a score so go back to the table selecting a score is nice and easy now my advice here is to write your own table i know the aig will provide a standard but they can't know your industry they can't know how easy it is to reference the severity of a problem write your own and try to make it clear and unambiguous but we have to give the severity a score it's usually from one to ten so i'm gonna give this a severity in this case i'm gonna give it a severity of eight i haven't got anything to relate it to so i can't you to the television industry i haven't got anything to relate it to so i'm high scrap rate for me would would be quite a a serious problem so i'm going to give it a 7. then it says what are the potential causes now when we go to the potential causes box here's something you must never write in the box and i know in the uk rolls-royce making aero engines for jet airliners they've told their supply chain this is something you cannot write in your fmeas and it's this you cannot write operator error in the list of potential causes operator error now then why do i want you to avoid the phrase operator error not because operators don't don't make mistakes but typically if you write operator error down the solution the control is very weak because what we tend to say is we'll train that person that doesn't fix the problem that person could be trained and they could take the day off the temporary person stands it they might make the mistake now or the trained person gets promoted and you forget to train the new person so operator error and operator training rolls royce in particular will not accept operator training as a control quite rightly because it then requires the operator to remember that standard and that training okay so what are the potential what are the potential causes for making them making the mistake so drying time too short what's the potential cause of making that mistake let's put something in the box so i'm gonna say um standard not visual to the operator and i'm also going to say and no tools to measure time all right so we've gone drying material drying time is critical to quality the drying time was the failure mode well the drying time could be too short yeah so we don't let the plastic dry what's the failure effect well if we use that in the process we'll get white marks in the molding what's the what's the potential cause well the standard is not visual to the operator in other words it's it's buried in a book or it's buried deep in a document somewhere how likely is this to happen now the occurrence now again this again has got a scale the occurrence and again it's normally it's normally one to ten but let me show you the one that i recommend in my um in my book drink tea and read the paper so in drink tea and read the paper the rule i use for how likely is it to happen is it likely to happen once a day once a week once a month once a year okay now i only give it a one to four scale now of course sorry but the other way around it four three two one you could go once every two years once every five years once every ten years if you really want to now why did i why do i pick this scale so it's back to what i'm saying before this has got to be easy to fill in one of the reasons fmeas take forever to fill out is because everybody argues about where we are on this scale now what i'm basically asking you to judge is how often are you going to upset a customer how often are we going to upset a customer are we likely up to upset the customer once a day once a week once a month once a year now normally in a conversation it'd be quite easy to say well yeah we probably wouldn't make that mistake every day because if we made it a couple of days running we'd soon realize that the dryer wasn't working properly but you know we might say well we're likely to make it once a month if we're not concentrating if the operator is not concentrating if we're rushing you know we don't have equipment to tell the operator how dry something is so even though they're trying to adhere to the rule there might have been a shift change and they might not know when the material was put in the dryer so they don't know whether the material is dry or not so we might make that mistake once a week we might make that mistake once a month and the idea is that it becomes very easy to say yep okay we'll pick a two yes okay we'll pick a three now in this case i'm gonna pick a three i think we if we if we're not telling people what the standard is and we don't give them tools to help them to judge how dry the material is how long it's been in the dryer i would expect us to make that mistake just by accident once a week so i'm going to give that a score of in this case three let's put that in there and then it says what are your your controls what are your planned controls so if he's if this is a new this is a new process this would be whoops spelt that wrong this would be what are your planned controls for this okay so we could say okay our plan controls specific we're gonna have a specific drying area okay we we are gonna have bins with clocks and alarms on so we're going to make the standard obvious or and we're also going to have so we're going to have a bin with a with a clock on it which you can then program but we're also going to have the standard displayed next to clock all right so there's our there's our control okay okay so there's my that's my control mechanism then you've got to ask the question okay how likely are we to d it says detect now we don't really want to detect the problem we want to prevent the problem how likely are we and that's why the new standard which is about error proofing okay how likely are we to prevent or detect the problem now again let's look at a current method of determining how likely are we to detect so let me look at one table that i've downloaded from the internet a standard that i've downloaded from the internet again this is a card industry example i've downloaded this straight from straight from google and here are the here's the detection and prevention criteria i mean look at this it says if you score five that is between a 36 and a 45 chance of detecting detecting the fault if you score an eight there is a 66 to 75 percent chance that you're gonna spot the fault now tell me how the hell are you gonna decide 65 to 75 percent are you going to decide 45 to 55 you've got no chance with a table like that you need to rewrite this now i'm going to show you one that i've rewritten i don't know how well it relates to the fmea that we do i'm going to take a look at it now but i suggest you write your own and you will see that it relates to the tool i've used so what i've done is i've said i've got a control mechanism how good do i think the control mechanism is so let's take a look at my list so here it is look one to ten we start with one and it links in very nicely to the iig new rules which is number one is a mistake proofing device is fitted to ensure complete prevention okay it says a mistake-proofing device is fitted to alarm out if the input goes out of spec to minimize the defects and isolate the affected part three says tpm routines are in place daily four says tpm routines are in place weekly five says spc is being used six says visual standards are out at the point of activity for regular audits now i would say that's roughly where i am i'm gonna put visual standards out at the point of activity for regular audits okay so then it says operators are trying to do the job correctly a skilled person will know what to do tpm is planned yearly finally there's no control at all now i think i'm around six at this point because i've got visual standards out at the point of activity so let's go to my table now i'm actually gonna score it slightly better than six i'm gonna give it a five why am i gonna give it a five well i've given it a five because i have more than just visual standards out at the point of activity i have a clock with an alarm so i have a sort of a mistake proofing device it doesn't tell them that they take they're making a mistake however it tells them they've hit the standard which is a slightly different thing um if if we linked the timer to the lid so that they couldn't lift the lid until the timer timed out i'd be scoring this as a one by the way because now they can't get at the material until the timing cycle has finished that would be a one but i've got a timer that a lot alarms out so i think i'm slightly better than a six but given that this still relates to the person following the rules they could cut across all the rules and violate them i'm gonna give this a five okay and that's where they are so if we go back to the form now and we look at those three scores we've scored severity likelihood and detection so what did we put i put seven uh i've put five here one have i put for likelihood i've put three if i multiply those three through what do i get i get a hundred and five now again key thing before you start this activity you have to decide what this number is going to mean now bear in mind by the way that this was only a one to four scale this was a one to ten scale so what's the worst i could get 10 times 10 is 100 so 400 was the worst i could get on my scoring system because i'd only got to want the four scale here for the likelihood so 400 is as worse he could as worse he could be so what you have to do is decide rules if you say i'm above 300 here's how we all behave if i'm above uh 200 here's our one i'll behave if i'm above 100 maybe here's how i'll behave so this 105 would either be telling me everything's okay or it would be telling me probably given that i'm towards the bottom of my scale i'm doing pretty well you should try and improve it a bit better and the the area i would go to is full mistake proofing with a lock on the lid with the timer if we do that this will change to a 1 and this whole number will collapse to 21. now we have total control if we think this is bad enough if we think the severity is bad enough we might want to take this down to 21. now what i'm going to do of course i'm going to set up the molding machine with a drying area with the standards on the bin with the special tying glocks on the bins that's what i'm gonna do next so what i've done now i've told everybody in my fmea what my process looks like what my potential risks are in terms of making defects and upsetting the customer so i've told them how important each variable is in my process times temperatures speeds injection pressures all of those numbers will all be in my fmea and i will say if they get violated this is how important they are and here is how i intend to control them and now i will design the process with the fme8 for the drain machine for the drying process i will design the process with password protection for all the programs on the molding machine etc and then when i go live this thing is going to go live bang defect free making piles and piles of money from day one and that is the proper way to do an fmea that's what fmeas are for that to design processes into now of course will i hold this up to scrutiny will i let a team of engineers look over my process design and say have i forgotten anything yes but it isn't a three-day fmea meeting just filling out a form talking about what scores to have for hours and hours on end with no value whatsoever this thing makes money it makes your process go live bang from day one making money then let people scrutinize it let it get it but get it better you know even better case you've missed anything but once you're ready what comes out the other end well what comes out the other end is a control plan so on that the end of this thing where i'm saying i'm going to use timed locks on the dryers that's a control technique that's a control plan right there in the planned control column in this column down here we have a control plan and if you ever get a problem with your molding process if you ever get a problem molding the battery cover what do you do they do is they generate control plans they generate a control plan that means you go live right first time and make piles of money they also generate a control plan so that if you get a problem it's auditable and your problem solving just takes a few minutes or a few hours because these things are fantastic if you do them right and you use them for process design fmea the power of getting it right first time

Channel: Paul Allen

Views: 3,387

Keywords: FMEA, Severity, Occurence, Prevention, Detection, Shewhart, Lean Six Sigma, Six Sigma, Taguchi, Risk, Control Plan, Gemba Academy, Quality HUB India, https://www.youtube.com/watch?v=s2HCrhNVfak, Simplilearn, MIT OpenCourseWare 2.85M subscribers, Amir H. Ghaseminejad, Saravanan Kuppusamy, https://www.youtube.com/watch?v=ouMnoLcK6uw, nptelhrd, SPC, 5s, Paul Akers, TPM, DOE, Fisher, Juran, Deming

Id: hGIRCAVXmKE

Length: 35min 50sec (2150 seconds)

Published: Fri Feb 05 2021