 Hello and welcome back to session 7 on Quality Control and Improvement using Minitab. So, we are trying to see tools that are used in quality for visualization, data visualization like that. So, some of the tools we have already discussed in previous sessions and today we will highlight some more additional techniques which are used in quality and which has an interface in Minitab. So, what we are doing in the last session is that we are discussing about cause and effect diagram. So, let us go to that and try to try to recap what we have seen last time. So, one of the thing that we mentioned over here in cause and effect diagram is that all are potential cause over here ok. So, there are 5 m's and 1 e's over here. So, environment is one of the another important aspects which is not shown over here. So, some of the causes and sub causes may not be present in certain scenarios, it depends on what type of problem we are addressing like that. Here we can see is that man, material, methods, machines, measurements, so personal means man over here. So, in this case, so 5 m's are present over here, but environment is not so prominent over here. So, there is no as such environmental influence which is creating defects in the tank like that. So, these are the what we are seeing over here, these are the causes and these are the and these within every type of cause what we can say is x over here and these are the sub causes what we can see defectives from supplier can lead to defects on the tank like that. So, this is basically why we can think of and these are the x 1, x 2's like that variables and within that sub variables we can think of. So, x 2, x 3 over here, so this we can think about x. So, within x there can be potential sub variables, so x 1, 1 like this, x 1, 2 like this, so x 1, 3. So, for a given cause can there can be sub causes also. So, many have as an option to add sub causes also, sub sub causes also like that ok. So, there are 5 m's and 1 e's like that. So, and all these causes what we are mentioning over here is basically we can think about these are potential cause that that means that there are evidence or either from empirical relationship what we have seen by scatter plots or something like that or maybe theoretically there are some relationship that exist like that from literature we can figure out those things like that. So, before we identify all potential causes much brainstorming is required to identify potential causes because we have to eliminate the causes or minimize the effect of the causes basically. So, that defects on the tank can be reduced, so which is the CTQ what we want to reduce and so how do we do it in Minitab? So, let me just illustrate in Minitab how we are doing that. So, for that I will take some examples which we have already hypothetical examples over here which is taken from book and here what we can see is that these are the dimensions. So, 5 m's and 1 e's over here and the sub causes for these are highlighted over here. So, this is in excel sheet what we can do is that we can just copy paste this one to Minitab worksheet. So, our worksheet that we are using was named as visual data like that and we will place it over here and so we will copy this one causes and sub causes like that. So, here we will just copy paste this one over here and whenever we have done that we are ready with the use of Minitab. So, what we can do is that we can just save this one save worksheet and replace the earlier one so that this data is saved like that. So, we want to replace yes I want to replace this one. So, this data gets saved over here and what we will do is that we will go to now we want to use draw this diagrammatically and try to see and so in this case what we go what we do is that stat over here and then go to quality tools and there is a cause and effect option over here click cause and effect over here and then there will be causes mentioned over. So, if you have not mentioned this one so it will be blank like this whenever you start this one and effect will also be blanked over here which you can title like that. So, you have to just click this one in personal personal means man so in this case and then you have to identify where you have saved that one which is related to man over here. So, I have saved in C33 columns like that so I will highlight that one then second one I will highlight the next one so machines. So, then I will highlight the third one materials where are the sub causes like that so methods. So, then again I will go to methods over here then measurement over here. So, I will I will highlight measurements over here and the final one is environmental issues that is considered over here. So, effect is let us say that defect is defective items that is being produced like that so you can just name it this is hypothetical. So, effect is y y category over here so we can say some forms of defects like that which is creating defective items like that. So, title you can always give let us say cause and effect diagram or something like that C and IE or something like that. So, I click ok over here and immediately what will happen is that you will get a cause and effect diagram with titles and all. So, if you see and maximize this one you will find that different categories are given over here. So, you can just change the font size over here and these are the categories that is already identified. So, immediately this is like a fish bone what you can see is that defective items are produced which is in red and these are the sub causes which is leading to this one. So, these sub causes we need to see and try to minimize the effects of this. So, if there is a problem with alloy we need to monitor and we need to experiment and try to figure out which is the best alloy that does not produce any defects like that. So, this comes under experimentation like that. So, we have to our objective is to create cause and effect diagram. So, I know what are the potential cause which is creating the failures and then we try to block those causes or minimize the effect of those causes like that. So, may have gives you an option to draw the cause and effect diagram which I wanted to illustrate over here, ok. So, then what we have is that so the cause and effect diagram. So, then another important tools sometimes we use whether it is quality management whether it is lean management like that. So, some of the things that comes into our mind is that process flow diagram. So, one important aspects is process flow diagram over here, ok. So, what is required is that to identify where the problem is what we do is that we try to draw the flow diagrams of the process. So, this is a hypothetical example where we are seeing that credit card processing, bill processing like that. So, how it goes to the client and billing is done and internally billing processing is done and then it goes to the client end like that. So, there can be mistakes and any of the sub processes like that. So, what you see is that this is information that is received then client folder is located over here. So, then updated client folder over here. So, these are some information sheets over here, buyer information sheets like that. So, then update on clients like that and send the billing statement like that. This is one of the hypothetical way there can be processes and sub processes like that. So, this can be process one, process two, process three like that. So, all our processes over here. So, these are sub processes that leads to final billing statement over here. So, any problem in the billing statement can be can happen and any of the sub processes like that due to any mistakes in the sub process like that. So, over here also we can we can just draw the cause and effect diagram for this and we can figure out what are the sub causes over here. So, maybe by ignorance we have done this and that is due to manpower or somebody who is supervising the things or who is coding the information like that in the database like that. So, that can happen like that. So, then there can be other reasons like over here methods, machines, materials, it may not be manufacturing, it may be service processes also. So, you can draw cause and effect diagram in various processes like that. Here it is inaccuracy in submission of the bills to the client like that, how much we can be accurate like that, what can go wrong basically those are the causes we want to block those causes like that. So, whenever I have a process flow diagram immediately what I can do is that what are the failure, failure chance where it can fail basically. So, those things are the causes and we want to block those causes. So, everywhere manufacturing service everywhere cause and effect diagram is required and also we make a rough flow diagram over here we try to draw a flow diagram. Minitab does not have any option to draw the flow diagram like that. So, you can explore other softwares where flow diagram can be drawn like that. So, maybe smart draw is one of the option where we can draw flow diagrams like that and because that makes a visual impact and immediately people try to identify with the within that n number of sub processes where something can go wrong and which is leading to inaccuracy in submission of the billing to the clients like that. So, so, Minitab interface we have seen that how to use cause and effect diagram. So, we will come to a different topics and we will try to see what are the other things we need to know primarily before we go into quality control and improvement. So, you see the data that I am representing over here a secondary data which is known as secondary data somebody else has generated that one I am just using those data those are known as secondary data like that. Primary data what I actually collect as a quality engineer or quality professionals like that from the process directly I rely on this. So, this is the primary data source like that and this is the most authentic data which is used for analysis any kind of analysis. So, we need primary data which is collected by me ok. Sometimes secondary data does not make much sense because the information that is required or related to a process may not be available in secondary data like that which may be of primary focus for my research or any other analysis or projects like that. So, we believe on primary data. So, whenever I am generating primary data can come from various sources actual observations, survey questionnaires like that and maybe experimentation what we talked about like design of experimentation which is basically done by me when I am trying to improve your process. So, systematic variation inducing variability and then trying to see what is the effect on why like that. So, on CTQs like that ok. So, these are primary information primary data information which is collected. So, always believe on primary data source for any quality control and improvement ok. Secondary data is not so much reliable when we are talking about quality control and improvement. So, you have to go to the process collect the data and information and then act on that data based on the data analysis you act on that. So, make some inference make some decision and act on that. So, that is the objective. So, it can be primary or secondary. So, we are talking about primary data source that on which analysis is done when we talk about quality control and improvement ok. Another important aspect is that whenever I have told about cause and effect diagram. So, we can we can think of various how to handle those causes because I want to minimize the effect of the causes or I want to eliminate those causes like that. So, so, whenever I have a single cause in that case very easy to solve this is not realistic one, but there can be only one cause which is creating defects like that. So, there can be one cause and if we have to go to the root cause over here. So, if you have to go to the root cause over here. So, what can be done is that we can use 5 y analysis 5 y 5 times like that and we come to the root cause like that. So, this is quite easy and this can be done. So, I ask a question why this is happening then subsequently I ask again why this is this is the reason why that is also happening like that. So, like this I come to the root cause and then I eliminate the cause like that or minimize the cause like that. So, this is impractical scenarios where only one cause is leading to defects in CTQs like that and there can be scenario when we have multiple causes, but the there may be they are independent basically. So, they can be independent causes which is basically independent. So, this type of scenario so, when they are independent. So, what happens is that there is little interference between the causes like that they are very independent like that and in that case what analysis quality quality tools is used over here is known as failure mode and effect analysis. So, if it is done in the design stage it is known as design failure mode and effect analysis and if it is done on the process stage when it is conformance stage when we are just implementing these those things and try to see the CTQ effect like that in the process. So, that is known as process failure mode and effect analysis. So, failure mode and effect analysis is to try to minimize the effect of the causes or maybe eliminate those causes occurrence of those causes like that. So, that is failure mode and effect analysis we will discuss something some briefly about this failure mode and effect analysis which is an important tool in quality. So, but you have to remember that all these causes what we are trying to address over here are mostly independent like that. So, one does not influence the other one like that. So, if I take corrective action on one that does not impact the other causes like that. So, or there is no relationship between cause 1, cause 2 and cause 3 or very minimal interface like that. But scenarios this is also sometimes impractical like that, but nowadays they are trying to make modular design. So, in that case it becomes easier to handle those things. So, in that case sometimes this design failure mode effect analysis is quite efficient to handle the or minimize the defects or defective items like that. So, this is a causes. So, scenario design scenario should be like this. So, all are independent whenever something goes wrong. So, I can just replace that one. So, immediately a modular type of designs like that. So, minimum interplay between the causes like that. So, that is the most favorable scenario, but scenario may not be like that where scenario may be most typically like this cause 1, cause 2, cause 3 and cause 4 where there is a fair amount of interaction that means one they together act on the wise like that. So, there is a interaction between cause 1, cause 2, cause 3 like that cause 4 and there is a complex scenario what we are seeing in this and then eliminating the cause becomes very difficult or minimizing the effect of the cause becomes very difficult. So, there is a interplay between the causes over here. So, there is a interplay between the x variables over here and that can impact the y over here. So, if that is the scenario when multiple x are interacting with each other. So, in this case scenario is not so simple, this needs to be addressed in a different way. So, when we when we are facing such kind of scenario mostly what we will see is that we need design of experiments for that to deal with that one. So, there are three ways we can deal with this one is y y analysis if the causes are single causes like that and we can address that one or maybe we can use failure mode and effect analysis when the causes are mostly independent like that. And the third one maybe when interplay is there we use design of experiments or statistical experimentation to minimize the effect of the cause like that or set the set the factor x where we are what we are mentioning over here in such a level so that we get the best delivery on y or CTQs like that ok. So, these are the different ways. So, let me try to just briefly talk about failure mode and effect analysis. What is failure mode and effect analysis? How it is used in design quality of design. So, if I have to improve the quality of design. So, in that case this is a typical example what was taken water data system for any residential homes like that. So, and so within the water data system there can be thermostat what you can see over here. So, there is a thermostat which controls the temperature basically ok. So, and it can fail it can also fail like that ok. So, then there will be a potential failure mode of this item or thermostat which is within the or component which is within the system water data systems like that. So, there can be multiple components like that and every components can fail every components can fail. So, if it fails that in which way it will fail basically potential failure mode this is potential failure mode over here. So, failure mode maybe it is not reacting. So, I am taking one example over here which is thermostat and fails to react to temperature rise like that. So, this is the potential failure mode ok. So, if this happens if this happens failure happens and it does not react to the temperature what we will have what is the effect on the system what is the effect on the system. So, temperature water temperature will rise water turns to steam this can be the two possibilities like that. So, one failure mode can lead to two effects like that. So, it is the effect one and effect two like this. So, this can be potential failure modes we can think of and then we can think of that there can be one effect and there can be two effects like that. So, there can be sub effects like this for a potential failure mode. So, failure mode and effect one and effect two like this. So, there can be multiple effects for a single failure mode like that ok. So, anyway, so we are talking about potential effect of the failure. So, whenever there is a failure mode and there is a effect on the whole system like that it will affect the system basically and whenever there is a failure it should be because of some causes over here. So, I have just identified hypothetically one of the cause non-functional thermostat basically. So, this can be. So, what I am trying to say is that there is a potential failure that can happen because of certain reasons like that and then which that reason may be of several causes that may be of several causes like that. So, one failure mode can be for various reasons why this has happened because of various various failure possibilities, failure causes various causes of causes are there which leads to such kind of failure basically. So, there can be multiple access like cause and effect diagram what we are saying. Say effect is basically a failure mode over here and x are the causes what we can think of over here. So, in this case cause 1 cause 2 up to cause n what we can see like that ok. And then every cause when we are trying to draw the Pareto what we have seen is that there can be different reasons for failures of the systems like that. So, some of the some of the reasons may be very frequently arising like that when I draw the Pareto which is appearing frequently and there can be scenarios like which is not occurring so frequently like that. So, cause every cause will have some frequency of occurrence like that ok. There is evidence that it will happen. So, because of this reason and that is why I am considering this in the in the failure mode and effect analysis. So, this is the basic format what we what you will find in failure mode and effect analysis. There will be item over here potential failure modes potential effects will be identified on the system and all the components like that and potential cause will be identified there can be multiple causes and for each cause what is the occurrence frequency that also that will also be noted down like that. So, is it very frequent is it not. So, here it is written remote possibility that is rarely this type of cause occurs like that. So, occurrence probability or something like that we can think of. So, cause occurrence over here if this is the cause is there any detection method that we have to detect failure because of this cause like that. So, detection method that is being followed over here and which will indicate that because this is a cause because of that. So, failure has happened like that. So, this is the reason basically. So, do I have a detection methods like that? So, is it in place? So, do we have that one or there is no detection method to identify this cause ok. So, and then recommended action in case it fails. So, what is the recommended design control that we have already mentioned already already in place in the design like that. So, these are the recommended control action. In case it fails and because of this cause then what action do we taken like that. So, what is important is that failure mode and severity of this which is the effect how much severe it is. So, that that that is also important how much severe it is and then what are the causes of that because if it is very severe we have to take action. We have to block those causes which is creating this failure mode basically ok. Then for every cause what is the frequency of that that also we we need to identify. Then what we can think of is there any detection method to identify if there is a failure because of this cause is there any detection method. Then accordingly we have a recommendation for that. So, this is what we do in basically a failure mode and effect analysis and there are different formats for this doing this one. So, if you go to NNHC we will find that design failure mode and effect analysis format will be there. So, potential failure mode you can see over here and items is mentioned over here. Potential effect of failure what I mentioned will be mentioned over here and severity of this failure that means this is very severe harmful to the society or it is mild effects that we have in the process like that or in the or in the or where I am using this particular environment where. So, severity in a scale of 1 to 10 they gives and there is a guideline to give the severity also ok. Then class of problem very critical like that. So, they may use some symbols like that if it is very critical. So, this type of failure mode is very critical like that. So, severity will be high basically. So, if the failure mode and effect is effects on the mankind is very high. So, in that case severity rating will be very high like that. So, these things are rated in a scale of 1 to 10 let us say severity rating what we are seeing and then potential causes over here cause will be written for a failure mode. So, one failure mode we have identified let us say and there can be because of c1 and c2, c1 occurrence rating will also be given over here. If it is very frequently occurring this type of cause we can give it a rating of 10. So, here also a rating of 1 to 10 will be used over here and then is there any way to detect these causes over here. So, again detection for every causes will be given over here. So, in this case so failure mode severity of this will be written over here severity one of this for failure mode. So, this is in a scale of 1 to 10 then occurrence of each of this cause will be given over here and detection. So, if I can detect immediately that rating will be less. So, if I can detect and but if I do not have in design control over here and I cannot detect those cause directly goes to the customer. So, in that case rating will be high. So, it may be rated as 10 over here. So, this is 10 severity occurrence also in a scale of 1 to 10 like that and this is 1 to 10 like that 0 is not included over here. So, because what we will do is that we calculated risk priority number out of this. So, this was started in 1949 I think approximately that time point US military force was using this one then NASA has used this one in 60s around 60s then automotive industry action groups who develop this AI AG which who develops also this QS 9000 quality system 9000 for automotive industry like ISO 9000 system they have developed initially and there they have given this format what you can see over here. So, this is this is the generic format that they have recommended over here and then they calculate a risk priority number for a given cause over here. So, the risk priority number will be calculated. So, I want to prioritize. So, what they will do is that they will multiply severity with occurrence with detection over here and that gives a risk priority number over here. So, for every cause there will be a risk priority number like that. So, risk priority number for cause 1, risk priority number for cause 2 like that. So, for a given failure mode for a given failure mode and then they will rank all this priority numbers like that. So, then based on that they will take corrective actions like that and redo the FMEA. So, every time you make some improvements, recalculate severity will not change, occurrence will reduce if you take some corrective action and detection will improve if you improve the detection in that case your score will go down. So, if the frequency of occurrence of the cause goes down the score will go down and if you have a strong detection over here. So, in this case we will go down also. Then a risk priority number will also go down like that. So, this is risk priority number, every company decide what should be the cutoff for risk priority number like that before they take the action. So, what they will do is that they will do a Pareto analysis of this risk priority number of each of this cause and they will do a Pareto over here. So, top risk priority number. So, maybe for cause one is very important cause two and cause three, we have risk priority numbers and then we will tackle this first because this is very critical and the risk priority number is very high maybe greater than 120 what we are getting. So, this we have to consider over here. So, generally risk priority number more than certain cutoff let us say 100 or 120 like that we will take corrective actions like that. So, this can be severity in a scale of 1 to 10. So, what I am telling over here, this will be also 1 to 10. So, maximum score you can get 10 multiplied by 10 multiplied by 10. But this does not happen generally we have some detection possibilities of this. So, in this case priority numbers can change. So, this. So, then severity how do we give severity ratings like that there are guidelines if it is very hazardous like that and appears without warning the effect is very hazardous and it will not give any indication to the customer. So, it will happen immediately like break failures or something like that it is hazardous like that. So, rating will be 10. So, severity rating will be 10 like that. And severity rating will not change for a potential failure mode. So, that is severity rating then occurrence how frequently this cause is coming like that. So, if you are taking if you have blocked that cause and occurrence you can reduce like that here also very high frequency of occurrence is very high this type of cause is coming several times. So, rating will be high. So, in this case and if detection method is not strong for this cause what will happen again ranking will be 10. So, so the minimum score is 1 and maximum score is 1000 what we can think of. So, this is 1000 maximum and minimum is 1. So, it cannot be any of this cannot be 0 like that. So, there is a risk priority number what we get out of that ok. So, this is design failure it can be done in design stage also it can be done in process stage also. So, whenever process in a place we want to improve that one and we want to block the cause. So, also we do process FME we do process FME. So, if you go to a company and quality control departments or quality assurance departments what you will find that they have a failure mode effect analysis and design if you go to a design where design department they are also you can find design failure mode effect analysis is being done. So, a general guideline is given if you want to see guideline is given in this QS 9000. So, so those things you can see around 1982 this was proposed like that. So, Ford, General Motors and Chrysler. So, they develop this one and they have developed given a framework to do failure mode and effect analysis in design stage and process stage like that. So, that can be followed as a guideline and this priority number can be prioritized and then accordingly which cost to be tackled first which cost to be taken second like that we can prioritize that one. How do you block the cost and how do you minimize the effect of the cost like that in the design or in the process like that ok. So, there is another important techniques that we will discuss in which is used in design basically is robust design the concept of robust design. What happens in robust design is basically when this was in 1980 when this came into highlight. So, Taguchi is one of the engineer who developed this concept of robust design and what was observed is that TV is manufacturing Japan Sony same company manufacturing Japan is for better than manufacturing in US. So, variability of the process when it is in Japan. So, you see variability is less over here as compared to which is manufactured in Sony. So, this is the specification let us say upper specification and this is the lower specification limit of the products like that. So, variation is too much over here when US products are developed. So, this is for a specific CTQ over here and in that case what was observed is that maybe picture clarity over here which is color density or something like that which was monitored at that time point and recorded. And that this is that the CTQ is variability is very high for US and for Japan it is very less like that people are buying Japan and because of this behavior what is what you can see. So, picture clarity it is always hitting the target value with minimum variability like that. So, that was happening and that is why so people are buying more TVs from Japan like that. So, then why this was happening analysis was done. So, in that case what was seen is that this noise variable is impacting basically outcomes of the process CTQs like that. So, then a concept of robust design came into in quality methodology like that. So, this was implemented in design stage also and in process stage like that. So, here interplay between these controllable variables and the uncontrollable variables what you see over here. So, interplay between this was used and still used nowadays also to get the best outputs like that although I cannot control the uncontrollable variables over here I do not have options to do that one. So, but what I can do is that I can always minimize the effect of that. So, that is the concept of robust design what we will discuss afterwards also. So, concept is hit the target and with minimum variability hit the target. So, mean and variance of the mean and variance both are important. So, this is in the design experimentation emphasize that both has to be mean should be in target and variability should be minimum and in one go I should do this. In one go I should not do it that first I adjust the mean and then I do the variability adjustment like that. So, let us find out setting conditions like that. Let us do it together and immediately one setting and with minimum experimentation let us figure out that variation will be less for the CTQ even in presence of noise variable which is uncontrollable variable like that that is known as robust design and this concept is very popular although it is controversial, but still people are using this in design phase you go to any car industry they may be implementing this one. This may be used in screening experimentation also before we implement we do full-fledged experimentation or response surface designs like that. So, this can be used for a screening factor screening also. So, that we will discuss afterwards at the end of the course ok. So, history says that history of quality management if you are just highlighting which are the key concepts that was developed at this which are the key things that has happened. So, 1923 approximately around that that time point Ronald Fisher developed this design of experiments. 1924 she wore developed the control charts 1930 these are the milestone achieved like that acceptance sampling I told that that cannot reduce variability, but that was developed 1930 and still people are using in processes for supplier products when they are supplying to the to your company what type of plans to you do you implement. So, acceptance sampling plan is there. So, ok. So, then 1950 around Ishakov I discussed about cause and effect diagram, 60 around zero defects concept crossbease zero defects concept came into existence. So, I want to minimize the defects like that 80 Taguchi what I told is the robust design concept was popularized at the time point although he started the work around 50. So, 87 ISO 9000 or which was taken from the concepts of total quality management that came into that gives a guideline to the industry and companies or organizations how do we implement quality philosophies like that in your organization like that and what are there will be some 8 principles based on which this ISO 9000 was built like that. So, 1987 8 principles of TQM so, that was used to develop this one. So, this is ISO 9000 and there are certifications. So, if you are a ISO certified company you some of the your client may be willing that may be willing to see that certification whether you are ISO certified or not, but many organization may not prefer to get this they may be needing some other certification like that. But this is in European organization mostly they asked for this ISO 9000 certification. So, this was revised in 1984 87 it was first started ISO 9000 1987 version then 1994 and currently maybe 2015 was the latest version like that. So, this is an international organization for standardization which has implemented this one. Then Six Sigma was developed 86 to 89 like that and lean Six Sigma philosophy also came into existence at that time point. So, Motorola started this concept of Six Sigma like that. So, we will discuss some aspects of this also and 2000 I told revision of ISO 9000 2008 to another revision came 2015 is the fifth edition of ISO 9000 that is I can these are the milestones what we can think of there can be many more like this, but I have just highlighted some of them. So, we will we will stop here discussion at this time point and we will start with control aspects from the subsequent classes. So, I hope you have understood the basic concept. So, failure mode and effect analysis you can see many books are there on failure mode effect analysis, but what is important is that severity occurrence and detection. So, that gives you a risk priority number, but that is that also I mentioned that this can only be used when we have causes which are very independent from other causes. So, then that is a key idea what I want to emphasize over here. So, cause and effect diagram is very important tool in quality and then Pareto diagram is also important to prioritize risk priority number what we have mentioned like that and based on that we will take some corrective recommended actions like that. So, then robust design is also used in quality of design like that. So, these are all things which you can think of in quality of design what generally people follows and some of the quality tools which are also used in this quality of design and then we will subsequently start with next sessions we will start with quality of conformance ok. So, let us stop over here and continue in our next class session 8 on quality of conformance with quality using Minitab ok. So, thank you for listening, we will again start in session 8. Thank you.