Quality Paper Final Tools Submission

TOTAL QUALITY MANAGEMENT

RESEARCH TOPIC The Seven Tools of Total Quality Management

Vikram Seshadri Student number : 0719005921 Professor : Prof. Harish Date of submission : September 30, 2007

Declaration Unless mentioned this report has not been submitted or been copied from any other work anywhere in the world and has been an individual research work.

Contents Introduction……………………………………………………………………….... 3 Why are tools so important?..........................................................................4 What is the various quality tools discussed in this report?....................... 5 Pareto charts……………………………………………………………………….. 6 Histograms …………………………………………………………………………. 9 Check sheets………………………………………………………………………..12 Scatter diagrams…………………………………………………………………...15 Flow charts………………………………………………………………………….17 Ishikawa diagram ………………………………………………………………….19 Control charts………………………………………………………………………21 Conclusion………………………………………………………………………….28 References………………………………………………………………………….29

2

Introduction

Total quality management is also known as management by data. It is highly improbable that a process can be stabilized and kept under control if there is no data available to analyze and verify the correctness of the procedures followed. In the present day world and working atmosphere, people maker more and more decisions based on only facts and especially in the western countries the decisions are no longer made by using basic intuition or gut feelings as we term it. Once we get past the small business concept, into the realms of a big business or a multinational company then data becomes critical and vital for controlling of the internal and external procedures and processes that they lead to. There are many ways to collect and analyze data that are required by the company. However carrying out this process of collecting enough data and then analyzing it can be tedious and a painful job in some situations. Hence, in this report I would like to look at some of the basic tools that are used in companies everywhere today and are extremely popular. These tools facilitate the process of collecting and then analyzing the data that we collect so as to arrive at a rational solution. Writing about the use of statistical methods in Japan, Dr Kaoru Ishikawa has said that “There are 7 indispensable tools of quality that are used by everyone: company presidents, company directors, middle management, foremen and line workers. These tools are also used in a variety of departments, not only in the manufacturing department but also in the departments of planning, design, marketing, purchasing and technology” (Goetsch, 2005). No matter where we work at what level in the organization, there is always going to be some or the other avenue by which we are going to use these tools like bread and butter in our work schedules. This report will take the reader through some commonly used quality tools and their interlinking and applications. It also deals with the involvement of the management in the process of control.

3

Why are tools so important? To explain this point lets think of an example. A construction worker when he works uses a few tools. He uses the drills, generators, planers, routers, sanders, saws, all power tools, hand tools, clamps, and hammers, all for different purposes but with a common mission and that is to make his job easier. All tools that we just read about are called as physical tools. These tools help him to work better and with more precision. However, total quality management tools are known as intellectual tools. These tools are used to simulate and improve human judgments. These tools are used to get better solutions and decisions in large firms and are applied in problem solving and decision making processes. A tool like a hammer exists to help do a job (Goetsch, 2005). If the job includes continuous improvement, problem solving or decision making then these tools fit the bill. Each of these tools is in some form or the other charts for the collection and display of data. Through the collection and display system of data, these data become important and can be used to solve problems and enhance decision-making. These tools can also be used to keep track of work done and can even help in predictions of future trends and problems. This would all but be impossible without these charts, given the mountain of data that one encounters in today’s work place (Goetsch, 2005).

4

What is the various quality tools discussed in this report? This report will actually handle an assortment of tools, some developed by quality experts whereas some are adaptive tools. They provide the means for making quality management decisions based on facts. It depends upon the situation, which tool will be used. Only one of them might be used or a combination of them might be used in some cases. There are a number of software’s available today to aid us in using these tools Total Quality Management (TQM) and Total Quality Control (TQC) literature make repeated mention of seven basic tools. Kaoru Ishikawa contends that 95% of a company's problems can be solved using these seven tools. The tools are designed for simplicity. Only one, control charts require any noteworthy working out. The tools are: • Flow Charts • Ishikawa Diagrams • Checklists • Pareto Charts • Histograms • Scatter grams • Control Charts

5

Pareto charts

The Pareto chart is a very useful chart, which is used in most of the industries today. It helps us separate the important data from the trivial bit of information. The Pareto chart is named after the Italian economist and sociologist Vilfredo Pareto (1848-1923) and was widely propagated and used by Dr. Joseph Juran. The Pareto principle is very simple. It recognizes the fact that in the world, in all processes, a majority of the issues that crop up are due to a minority of causes. This concept can be understood better with the help of a simple example. Consider in a factory, if we analyze the situation we find that 80% of the defects are due to 20% of the problems that exist in the factory. The remaining 20% of the product defects are caused due to the remaining 80% of the problems that exist. Examining the cost defects, we can analyze and see that 80% of the total defect costs will be raised thanks to only 20% of the cost elements. This fact is true in any other industry, considering the classic case of a classroom; we would find that 80% of the problems caused in the school would be due to only 20% of the students. The remaining 20% of the problems is caused by the other 80% of the students. An example of this case is given below in the form of a Pareto chart, which the author has himself designed and conceptualized. In the chart below, we have two sets of data. One is the total number of children, which are 200 and the distribution of the characteristics of these children. We can observe from the chart that nearly 160 students are falling into just two categories of the chart that is the talkative and dull groups. 160 students comprise 80% of the total strength but two categories are just 20% of the total categories that the chart has. This shows us and reinforces the fact that 80% of the defects are caused by just 20% of the problems. If we can handle our resources in such a manner that we

6

concentrate on eliminating the talkative and dull behavior in class then we would have solved the problems of 80% of the students.

Number of students versus problems in schools

Talkative

dull

weak In studies

naughty

others

Therefore, what a Pareto chart does is that it shows us our main reasons for concern or the main areas wherein we need to focus most of our resources to get corrective action carried out immediately. It helps us prioritize our resource expenditure in a very orderly manner. Coming to another variety of Pareto charts, we need to talk about the cascading Pareto charts. Say in case we do successfully find out that “too talkative students” is the main problem then we need to get to solving that. Therefore, what we do is we take this one factor, draw a separate Pareto chart on its 7

elements, and divide this one factor into many small factors so that we can allocate our resources more carefully and precisely. For example, take a look below at the Pareto chart that has been drawn and observe that if that one factor, “too talkative students” is divided, and then it will still follow the 80-20 rule.

As we can infer from the diagram above the 90 students who were the major problem are now divided into smaller groups to find out what are the main causes of them been so talkative. In addition, from the Pareto we can make out that almost 80% of the students are again from just two problems that are 20% that is they are just too active or are just very clever. Hence, in this section we saw the uses of Pareto and its various applications. Now let us see the next tool that is used very commonly and that is the “Histograms.”

8

Histograms The first statistical process control tool that is usually used is the histogram. The histogram graphically estimates the variation and identifies the gaps in a process. When we have a set of process data, a histogram maybe used to graphically display the process and its variables. A histogram can be built by dividing the range of the data into equal sized portions. For example, if your data ranges from 1 to 7, you could have equal partitions of 0.5 consisting of 1 to 1.5, 1.5 to 2.0, 2.0 to 2.5, and 2.5 to 3.0 and so on. The procedure is very simple. Collect at least 50 samples for plotting and then calculate the range of the values, which is the largest value minus the smallest value. Then determine the number of classes for the X-axis and then calculate the size of each class. The class interval can be obtained by dividing the range by number of classes. Note down the data class lines (the maximum and minimum values that can be plotted in any class value). Finally calculate the frequency of each class and plot the histogram. The histogram would answer various questions such as what is the shape of the distribution, what the most favorable response is given by the system, does the data contain outside restriction points, does the data take a symmetric form or is it skewed to the sides and so on. What we do is we take the frequency values that are tabulated and then we display them graphically. The major difference of a histogram from a bar chart is that in a histogram it is the area of the bar that gives us the value and not the height where as in a bar graph it is the height of the bar that gives us the value. This is very important for us when the categories are not of uniform width (Lancaster, 1974). Variation in a pattern always exists and this variation is usually captured by a histogram. Histograms provide clues about the characteristics of the population from which its sample is taken. While a fixed and clearer picture can be seen by using a histogram, it is also important to note that the same data in numerical 9

form or in a table would have been very hard to understand and comprehend (Evans and Dean, 2004). An example of a histogram is shown below.

The mathematical explanation of a histogram can be given by stating three terms. The first one is ‘n’ which is the number of observations. The second is ‘k’, which is the number of bins as they are called (also known as samples). The third term and the final one is ‘mi’, which is the mapping of the histogram. These three factors are interrelated so as to give us a formula as below

10

An additive histogram is a plot that counts the summation number of observations present in all of the bars until the present bar. That is, the cumulative histogram Mi of a histogram mi is defined as: (Besterfield, 2005).

(Besterfield, 2005)

Now that we have taken a look at how the histogram works lets take a look at the next quality tool and that is one of the most commonly used tools in quality, namely the ‘Check Sheet’.

11

Check Sheet

Many companies in today's world have a rich access to data and with the advent of powerful desktop and palmtop computers are able to collect much information. Nevertheless, sometimes this information itself is too much to comprehend and can actually drown the company in confusion. In such cases, it is important for a company to distinguish and segregate the useful data from the trivial data available. Check sheets help us do exactly that. The check sheet is a valuable tool that can be used for a series of applications. Its only restriction is the imagination of the person noting down the data. In addition, one more important advantage of a check sheet is that it can give us the data that we can use to prepare more complex tools like Pareto diagrams and so on. There are four different kinds or ways in which we can prepare check sheets and they are –

•

Defective item check sheet –

In this kind of check sheet, we look for a kind of defect in a product. In this check sheet usually there will already be a list of defects and as and when we check the samples if in case we do find a defect then we put a check against that defect. This is a countable kind of data check sheet where the results can be counted and tabulated (http://class.et.byu.edu/mfg340/lessons/, referred on 22 September 2007). In the following page, the author has given an example of the defective item check sheet used in a sand casting process. It can be observed that every time a defect is observed, a check or a cross is put in the column. The total number of checks will give us the variation.

12

Types

of

defects

Frequency

Cracks

xxxxxxxx

Blow holes

xxxxx

Porous

Hard finish /

castings

brittle

xx xx xx xx xxx xxxxx

Total

•

8

5

13

3

Defective location check sheet –

This kind of check sheet is usually used in processes wherein the exterior finish of the product is vital. Usually these check sheets have pictures or diagrams of the product and checks are put on these pictures to indicate where exactly the defect is occurring.

•

Defective cause check sheet –

This kind of check sheet takes into account the reasons or causes for a defect to occur. It can actually be used to measure and compare two or more variables through the same check sheet. For example, consider the check sheet given below. Machine 2 has more defects running in the afternoon session than the morning session; hence, it needs to be looked at.

Operator A

Operator B

Machine 1

Machine 2

Morning

X

X

Afternoon

XX

XXXXXX

Morning

X

XX

Afternoon

XX

XXXXXXXXX

13

•

Checkup confirmation check sheets –

This is the most common kind of check sheets used in industries today. These check sheets usually have a list of jobs to be done and as these jobs are completed one by one there are going to be a check mark put against it. This kind of check sheets are usually used in maintenance schedules and service industries.

14

Scatter Diagrams This is one of the simplest and most common quality tools used. It is nonmathematical and is so simple that anyone can do it. In scatter diagrams, we take two variables, which are dependent on each other and plot their values with respect to each other. This plot will then give us the kind of interrelationship that these two factors share. Depending upon the kind of interrelationship that the two variables can share, there are three different kinds of scatter diagrams and they are – •

Positive relationship scatter diagram

•

Negative relationship scatter diagram

•

No relationship scatter diagram

All the three types of scatter diagrams have been shown below in fig 1, fig 2, and fig 3 respectively.

Fig 1: positive relationship

15

Fig 2: negative relationship

Fig 3: no relationship

As you can see from the above three graphs X is the driving factor such as time of boiling or heating and Y is the performance factor like the temperature and so on which depend on X in some cases and do not depend on it in some cases. This was about check sheets. Now let us see the next quality tool and that is the flow charts.

16

Flow charts For many processes which have a number of sub processes present within them, we can have a process flow diagram, which will give us who the next customer of the process is. These diagrams facilitate the understanding of the entire process and what should be the guidelines that need to be followed. Let us consider the example of an incoming call router and its process flow diagram.

(Source: http://www.mindtools.com/media/Diagrams/flowchart.GIF, referred on 20th September 2007).

17

At the beginning of the call we need to answer the call and ask them how can we help the caller. If the caller has a problem, we enquire about the problem and then we take down the company name and the callers name and transfer to help desk and then that stream of the process ends. Another stream that the same process can take is if the caller wants to place a new order or wants help doing it then we transfer the call from the second stream itself that is we do not go ahead to ask them if they have a problem and so on. In today's world many companies such as the Coca Cola company, Colgate Palmolive , Columbia pictures and Antz corporation use the flow chart extensively to explain their company processes

and procedures to new

employees and changes in it to current employees. The advantage with this tool is its simplicity and its ease of use. It can be understood by anyone and can always give the correct interpretation of the process. What’s more, since everyone can understand the process so well it is open for improvements at any given time by any employee. Now that we have understood how a flow chart works, let us look at how the next quality tool works and that is the cause and effect diagram or the Ishikawa diagram.

18

Cause and Effect Diagrams/Ishikawa Diagrams The cause and effect diagram is the most commonly used analysis tool to find out the main reasons for the final effect that has happened. A cause and effect diagram is also known as the fish bone diagram or the Ishikawa diagram named after the founder and promoter Dr. Kaoru Ishikawa.

The cause and effect

diagram is used to either get to know the reasons for the bad effect or the causes for the good effects. Causes might be many and for ease will usually be broken into six categories of man, material, methods, machines, measurement and environment. There might be other cause sectors in industries like the service industry, which can also be put up. Each major cause is further divided in minor causes. For example under methods we might have ability, training, physical characteristics and so on. For explaining and visually understanding, the cause and effect diagram better, let us look at the illustration given below

19

The above fish bone diagram is for a quality problem in a sand casting industry. We have taken the six major causes as work force, machines, environment, measurement, working methods and materials. These major causes are again divided into smaller minor causes such as arrogant managers, demotivated workers, and pressure to perform improper calibrations and so on. All these causes lead to the one bad effect and that is the hard and porous casting. Finally, the author would like to say that even though the cause and effect diagram is a very simple diagram yet very effective as it gives us a diverse view and also fosters teamwork and can actually be used in any issue arising n the business and it can be fun. Now that we have had a view of six of the quality tools that are used let us look at the last and most important quality tool and that is the control charts.

20

Control charts This is the last of the tools but certainly one of the most widely used and most important tools in the workplace today. The development of total quality management by taking into account the usage of statistics and theories has been well documented and infact this has been the cornerstone for an improvement in any process. Using statistical process control an organization can continuously monitor, control and analyze the process so that they can detect any variation and correct it early in the process. The meaning of variation in a process is the degree to which there are errors moving away from the control limits specified or from the nominal value specified. Variation always affects the output of the process largely. Variation in a process may be due to special causes or local causes/system causes. Local causes/system causes are the ones that regularly cause a variation to occur. Examples of system causes would be the improper training of staff or workers, poor design of workplace, supplier credibility problems, and so on. Special causes are those causes, which usually tend to happen in a periodic manner and are mostly not expected when they happen. These causes are usually dealt with by the machine operator himself or by the net level worker in the organization. Some examples would be material slipping, machine breakdown, oil leakage, and tool broken and so on. What we can see from the above two causes is that special causes are the ones that do not always cause a worry for production as such things happen. However, it is the system causes, which the management should look at as these causes are the ones, which usually lack knowledge about and will continue to produce defects if left unattended. Infact Dr. Deming rightly noted that it is the system causes which are the root to 80 -85% of the process variations whereas special causes would only contribute to around 15-20% of the actual variation in the process.

21

The control charts were developed by Dr. Walter Shewhart in the early nineteenth century. Today it is the most widely used quality tools largely. The major advantage of a control chart is not only does it indicate variation but also shows us where the variation occurred so that we can immediately attend to the problem area. In order to understand and plot a control chart a user might have to have a basic understanding of how the math works the chart, but the chart can easily be interpreted by any worker in the organization. Today in the organization, we use mainly two kinds of control charts and they are the r-chart (or the range chart) and the average (or x-bar chart). The average chart or x-bar chart gives us the total variation of the sample points from the actual mean line or mean value that we designate. The r-chart gives us the range at which these sample points are distributed around the nominal mean. Let us first look at the various components in a control chart. There are three main lines and these are the upper control limit line (UCL), the lower control limit line (LCL), and finally the control line or the mean line (CL). In order to plot a control chart we need to take into account at least 30-50 sample values, which will be represented in the form of a point on the control chart. In order to understand the control charts better, we shall look at the example given below and as to how we can construct a control chart from the data given. This example will also show us as to how we need to measure and plot variation in a process. In this process we consider a tool room operator taking measurements of a carton box being manufactured. He takes a sample size of four that is n = 4 and checks on the measurements every hour. The main measurement that he is focusing on is the thickness of the carton box and its variation in the production process. Let us see how to plot a x-bar chart and rchart from the data collected.

22

First we need to tabulate our values, which we get by taking the samples out for measurement.

Sample 1

Sample 2

Sample 3

Sample 4

x-bar

or Range or

average

‘r’

8am

1.0

1.09

1.02

1.02

1.0325

0.09

9am

1.01

1.03

1.04

1.01

1.0225

0.03

10am

1.03

1.06

1.03

1.01

1.0325

0.05

11am

1.0

1.06

1.04

1.05

1.0375

0.06

12pm

1.08

1.02

1.05

1.07

1.0550

0.06

1pm

1.05

1.08

1.03

1.02

1.0450

0.06

2pm

1.05

1.06

1.03

1.07

1.0525

0.04

3pm

1.08

1.03

1.03

1.05

1.0475

0.05

4pm

1.04

1.05

1.03

1.06

1.0450

0.03

5pm

1.06

1.04

1.03

1.07

1.0500

0.04

6pm

1.0

1.08

1.04

1.08

1.0500

0.08

7pm

1.06

1.07

1.03

1.04

1.0500

0.04

X- Bar is just the average of the value of the four sample point values noted down. That is X-bar = sum of all 4 values of sample points Number of samples taken (n) In addition, range ‘r’ is the difference between the largest value and the smallest value of the sample points taken. That is Range or ‘r’ = highest value out of ‘n’ sample points – smallest value out of the sample points

23

After we calculate the range and x-bar for all the sample batches then we calculate xdouble bar, which is the average of all the x-bars calculated. That is,

X-double bar = sum of all x-bars calculated earlier Number of sample batches taken Therefore, in this case X-double bar = 12.52 12 Therefore, X-double bar = 1.043333 which can be rounded off to 1.0434. R-bar is the next step to be calculated and it is the average of the values of ranges that we have calculated. That is R-bar = sum of all the ranges calculated Number of sample batches taken In this example the value of r-bar is, R-bar = 0.63

= 0.0525

12 Now the next calculations are as follows, For the R-chart, the centre line will be R-bar that is 0.0525 Upper control limit or UCL = D4*(R-Bar) = 2.283*0.0525 = 0.1198 Lower control limit or LCL = D3*(R-Bar) = Zero*0.0525 = Zero

24

For the X-bar chart, the centre line would be X-double bar that is 1.0434 Upper control limit or UCL = X-double bar + A2*(R-bar) =1.0434+ (0.729*0.0525) = 1.0870 Lower control limit or LCL = X-double bar - A2*(R-bar) =1.0434- (0.729*0.0525) = 1.0051 D3, D4, and A2 are constants that are mentioned in any quality handbook and are constant for a given sample size around the world. Now with all these values calculated we can go ahead and draw our control charts. First let us draw the Xbar chart.

1.0870= UCL 1.0870

1.0434= CL

1.0051 1.0051= LCL

1

2

3

4

5

6

7

8

9

10

11

12

Sample number

25

Now that we have seen the plotting of an X-bar chart, let us look at how to plot an r –chart or range chart. Using the data calculated above,

0.1198= UCL 0.1198

0.0525= CL

0.0000 0.0000= LCL

1

2

3

4

5

6

7

8

9

10

11

12

Sample number

What we could observe from both the graphs is the fact that none of the points taken are crossing the control limits. What ever we can make out from the chart is that there is a pattern of flow among the points. The process is said to be under control and under the influence of only system causes if all the points lie within control limits. If there are any points outside the control limits, then the

26

process is said to be under the influence of special causes and the process needs to be verified and the special cause should be eliminated. Continuous improvement of a process can be achieved by control charts only if the process is under control first. When we observe the control charts in factories and other industries we can observe various point patterns among them and a few of them are runs, trends, cycles, jumps, hugging points. Cycles are said to occur when there are a number of sample points following the same value or pattern of movement. Jumps are said to occur when there is a huge shift between the location of one point and the next point. Hugging points are when there are two or more points that are very close to the UCL or LCL or even the control line. Trends are a dangerous situation that indicate tool or machine damage and this can be perceived to be happening when there are many points occurring on one side of the CL and also sharply falling and rising on one side only. Finally runs are said to occur when there are a number of points occurring on the same side (http://class.et.byu.edu/mfg340/lessons, referred on 20 September 2007).

27

Conclusion Quality is the responsibility of not just one person in the organization, but the responsibility of each member of the organization. To ensure that quality products are coming out the management has to be actively involved in propagating quality control and its concepts. When there is an abundant quantity of quality data available, companies can some times go wrong in analyzing what is not required and leaving out what was of utmost importance to the process. For example, consider a bank where there are thousands of customers carrying out different kinds of transactions at any given time all around the world. The data that is collected is scary to any one individual. Nevertheless, when a problem arises the bank should be in a position to collect the relevant data and analyze them to arrive at the solution at the earliest. Thanks to the efforts of pioneers of quality like Juran, Taguchi, ishikawa, Deming and so many others, we have a few tools in the world today that reduces the burden on us so much. The author just hopes that the information found in this report will help the reader get a clear understanding of the various tools of quality and their degree of helpfulness in case of a problem. A few more tools need special mention like the Deming’s cycle, house of quality, failure mode effective analysis and so on. These concepts are not tools but institutions of quality by themselves. Hence, these topics have not been handled in this report. Again hoping and wishing that these simple tools could help any aspiring quality worker in his/her job to attain quality work. Thank you.

28

References •

http://elsmar.com

•

www.isixsigma.com/tt

•

mrvar.fdv.uni-lj.si/pub/mz

•

deming.eng.clemson.edu

•

www.taproot.com

•

Ginevan M, Splitstone D, 2003, Statistical tools for environmental quality control, 3rd edition, CRC press.

•

Edosomwan J, 2004, Integrating productivity and quality management, 2nd edition, CRC Press.

•

Evans R James and Dean W James, 2003, Total Quality Management strategy and organization, 3rd edition, Thomson South Western.

•

Goetsch L David, 2005, Quality Management, 5th edition, Pearson International Edition.

•

Tsiotras, George, Gotzamani, Katerina, 1996, The International Journal of Quality and Reliability Management, Vol 12, Bradford publications.

•

Marsh S A, 1993, the key to TQM and world class competitiveness, Part 2.

29

•

Rogers E Rolf, 1997, Implementation of Total Quality Management, The Haworthe press.

•

Kanji K Gopal, Asher M, 1996, 100 Ways to Achieve TQM, Sage publications.

•

Ebert J Ronald, 2000, Business essentials, Prentice Hall.

•

Dettmer H Williams, 1998, Breaking the constraints for world class performance, ASQ Quality Press.

•

http://www.asq.org

•

www.businessballs.com.

30