Lean and six sigma

Understanding Lean and Six sigma and my Green belt Journey

2

Overview

I complied this Presentation to explain the differences between lean and six sigma to my colleagues within my organisation.

From there the presentation goes on to detail the project I led to achieve my Green belt certification in March 2012 and how this structured approach has changed my thought process towards problem solving within technical processes

My main development has been within six sigma rather than lean and for my next role I would ideally like utilise my existing skills and develop into a competent Lean Six Sigma Black Belt where I can drive sustainable improvement and help develop others

2

Overview

Stability and Accuracy Speed

3

Lean Focus

9

House of Lean

Vision Goal Philosophy Metric Method Tool SymbolBenchmark Value

6

• A level of performance that reflects the number of concerns in our products or services

• A statistical measurement of our process capability, as well as a benchmark for comparison

• A set of basic statistical “tools” to help us measure, analyse, improve, and control our processes

• A commitment to Customer Satisfaction and Shareholder Value to achieve an acceptable level of performance

Sigma is a letterin the Greek Alphabet

What is 6 Sigma

6-Sigma and process capability

7

Sigma Level DPMO % Good Cpk

1 691462 30.9 0.33

2 308530 69.1 0.67

3 66807 93.3 1

4 6210 99.4 1.33

5 233 99.98 1.67

6 3.4 99.99966 2

6-Sigma Strategy

Know What’s Important to the Customer

Center Around the Target

Minimise Variation

Reduce Concerns

8

9

•What problem would you like to fix? •create a Project Charter, •Create a high-level view of the process, • Begin to understand the needs of the customers of the process.

•Determine baseline of the process • look for clues to understand the root cause of the process (fill the funnel)

•What does your data tell you?•Narrow down and verify the root causes

•How will you fix the problem? •Move on to solution through hypothesis testing

•How do you sustain the newly achieved improvement? •Document exactly how you will achieve sustained improvementProject Progress

30 - 50

10 - 15

8 - 10

4-8

3-6

Number ofInput variables

KPIV’s to optimiseand control

The Six Sigma Methodology

The Six Sigma Tools

10

• Establish aTeam

• Identify a Sponsor

• Administer Pre- Work

• Pareto• Project Charter• VOC & CTQs

• Define current state

• SIPOC• Map the

process• Value

Analysis• MSA• Collect data• Check

sheets• Prioritization

using Pareto Charts

• FMEA• Control

Charts• Boxplot• Time series• Control

Charts• Histogram

• Process Capability •• Cause & Effect• Pareto• Scatter Diagram• Histogram• Boxplot• Multi-vari Studies• Interaction Plots• Regression• Analysis of

Variation• Hypothesis

Testing• FMEA

Brainstorming• Design of

Experiments• Pareto• Hypothesis Testing ••••••

BoxplotFMEAProcess Mapping Selection Matrix Pilot

• Control plan• Specification• Control Charts• Error Proofing• SOPs• Training Plan

Overall Approach

Practical Problem Statistical Problem

Statistical SolutionPractical Solution

y f ( x1 , x2 ,..., xk )

26

My Practical exampleDefining the Problem

Project Description: Realise £750,000 of raw material savings through rice pellet supplier change

Problem Statement: New Supplier product gives poor quality product quality compared to existing

supplier

Project Objective: Technical match of product of new supplier to existing supplier

In Scope: Supplier manufacturing process and popping process at Skelmersdale

Area Out of Scope: Seasoning and Packaging processes

Primary Metric KPI: Appearance defect %

KPI Definition: Technical match for appearance and popped cake characteristics

Secondary Metric KPI: Area Weaklink %

KPI Definition: Site Weaklink standards

13

D M A I C S

14

Product produced using new supplier rice pellets were found to give very variable product quality during trials. The quality of product was not consistent and failed to match existing supply. For this reason the trial was cancelled before completion. At this stage I was not aware if this quality issue lay with the site processes or with the proposed new suppliers process. During the trial I obtained the suppliers process data to help define were the focus areas lay and found the following.

A probability plot showed the proposed new supplier finished pellet process data to be non normal with data stacked at the specification limits and a capability analysis gave a Pp 0.38 and a Ppk of 0.19 indicating issues with process control.

When we measured pellet characteristics such as 100 piece weight we could see on the control chart that a lot of variation was seen from bag to bag showing a lack of consistency and control

The Supplier ‘s Process D M A I C S

Define actions

• Using the suppliers own data I could see that the process was non normal which, in this circumstance, indicated that process data was manipulated. The capability showed that the process was not capable within the defined specification limits. This suggest that the supplier was having issues controlling he process which was in turn leading to quality issues at the process.

• A quality assessment of the trial product and existing supply showed that some LT product matched current Liven supply where as other product was significantly worse and failed to meet on line quality standards.

• Through our process understanding we were not convinced that finished pellet moisture was the critical variable that affected popped product quality.

Our sister plant in the USA pellets for an identical process.

Actions• Proposed supplier process control needed to be improved through constructive feedback of

process analysis• Compare Current supplier against proposed supplier• Define critical variables within the process that affected quality• Use current business knowledge to build our understanding of pellet manufacture and critical

control points of the process.

D M A I C S

15

16

Comparing the current and proposed rice supplies it showed there was no difference in the variation and mean average for the finished moistures . When comparing bag to bag moisture of the proposed suppliers product against appearance defects results it was found that the bag to bag moisture could be the same for both good and bad product

To confirm the hypothesis that finished pellet moisture was not the critical variable I ran a fitted line plot to of Moisture against defects and found that there was not correlation

We created a SIPOC or process map to define the scope and some areas for review.

Good product

Bad product

D M A I C S

Measuring the Current Process

• The proposed supplier process was being controlled by finished pellet moisture as the primary control measure which was measured at the end of the process. This was shown to be very variable and could not meet specification limits. It was agreed that finished pellet moisture was the best way to measure control throughout the process and could be used to compare the process as improvements were made as the project progressed.

• On discussion with colleagues in the USA it was clear that moisture measurement throughout the process was critical for good process control. The proposed supplier was asked to measure moisture throughout the process and define specification limits for these to allow these variables to be measured. The critical control points were also defined through the process map of LT’s process.

• Extruder control was also highlighted as critical to improve LT baseline performance and was agreed as an area of focus and measurement area. LT were advised to start measurement of this process area and feedback results for measurement.

17

D M A I C S

18

Comparing the two extrusion lines at the proposed supplier we could see that line 1 was showing less variability and was statistically better in a number key areas of control at the extruder. This was highlighted to LT and they consulted the extruder manufacturer to improve line 2 to the same standard

Although line 1 was less variable the control charts for pellet characteristics of weight and diameter showed the process on both lines was not stable. More control was needed at the extruder to reduce the variation that was impacting size and diameter

Moisture control needed to be improved. Inconsistency in raw material in feed moistures was not being controlled at the pre mixing stage . Poor extruder control was also leading to variable post extruder moistures. Improvements in control in these areas will lead to reduced variability of the finished product.

D M A I C S

Measuring the Current Process

• After we feedback the findings of the variability seen in finished pellet moisture and statistical differences between the supplier processes they reported a number of issues that were affecting consistent running of the line were previously no issues had been reported. To rectify these they complied and completed the following improvement list.

• Replace flour feed hopper so that flour bridging is eliminated. • Add level probe on flour feed hopper to permit auto feed from mixing hopper to feed hopper• Add a water chiller to reduce cooling water used for cooling extruder to regulate zone

temperatures in the extruder. • Improve cam design on vibratory conveyors from extruders to make it more robust. • Replace electrical heaters in Vibratory dryers. • Obtain uniform bed in apron dryer to obtain more uniform moisture across the bed and to

reduce clumping:• Improve distribution using doctor blade in dryer feed hopper. • Improve controllability of inlet temperature in dryer. Reduce temperature fluctuation to +/- 2

Deg C.

19

D M A I C S

Analysing The Root Cause

20

D M A I C S

Once the lack of a relationship between finished pellet moisture was proven we focused our attention on gelatinisation of the pellet. This involved analysis of the pellets at a third party lab to test for uncooked starch levels using DSC and also pellet analysis using RVA.

I also visited a reputable University to build an understanding of starch and how is reacts within processes

The practical question was how does starch in the pellet affect the popped cake quality?

21

DSC testing shows the amount of uncooked starch remaining as an enthalpy value. This value was compared to the appearance defects observed on trial product. A moderate negative correlation was observed between the two. the hypothesis we derived from this was that the high enthalpy value = good product and low enthalpy value = bad product

As we analysed more trial material we found this hypothesis to be untrue. We found that we could get good product at low or even 0 enthalpy values! although we had not shown that high enthalpy values gave bad product

The next hypothesis was around extrusion and the energy used to work the product. The specific mechanical energy (SME) exerted on the product could alter the state of the starch granule causing quality issues (See picture above)

D M A I C S

Desired granule structure post LT extruder

Cell structure of poor product through increased SME and/ or not enough available water?

You need to be able to show this

A visit to the University was instrumental in building understanding of the starch granule and the right characteristics required for the popping processWe took the findings and complied the following guidelines for the proposed supplier.

Good Product: Right amount of SME and available water to swell the starch granule without deconstructing the granule. This allows good expansion at the popping machine and the ability to hold its form when popped. This is because there is a strong cell wall intact to maintain the popped shapeBad Product: Too Much SME with enough available water will deconstruct the starch too much giving good expansion but will not hold its form as the granule structure has a weak cell wall that will not hold as well.Bad Product: Too little SME will not swell the granule giving poor expansion as it will not take up the available waterBad Product: Incorrect amount of available water will give the same results as too much or too little SME

To prove the hypothesis we asked LT to reduce the work that done in the extruder and to send samples to the UK for popping

D M A I C S

22

D M A I C S

Low GelatinisationNormal Gelatinisation

Defect Rate = 0% Defect Rate = 2% • The hypothesis proved to be true and we observed very good expansion and retention of form on

the low gelatinisation (Or reduced mechanical energy) pellets .• On the normal gelatinisation we saw less expansion or retention of form, but still good quality

product in line with the current supplier• To achieve this the proposed supplier reduced SME in the extruder by reduction in heating zone

temperatures by 25oC on zone 2 and 15oC on Zone 3• Reduced SME has maintained granule structure and increased post extruder moisture by 2%

23

Improving the Process

24

D M A I C S

The improve phase was focused around defining critical control points for the proposed suppliers process and implementing measures for these areas as follows:

• Measurement of rice flour moisture pre production is critical• Correct addition of water to flour to maintain consistent 30% moisture within rice/ water mix is

critical. The supplier was currently adding the same water regardless of incoming flour moisture as seen in gelatinisation trials

• Determine optimum process controls a measurement of SME within the extruder would be very useful. SME combines all the factors within the extruder that will affect starch gelatinisation and granule structure.

• Best product observed at 30% infeed moisture with 25.4% post extruder moisture at reduced temperatures in the extruder. Define settings for process control.

• Checking of post extrusion moisture is a critical process control to ensure consistency in extruded product.

• Determining optimum gelatinisation level as measured by KOH is critical and is a required process control check. Best product observed (Low Gel) was KOH = 2, Normal was KOH =5 which gave very variable product

25

Defining the rice flour mix moisture as a critical control point gave consistency of product into the process

Understanding the reactions within the extruder and their affect on product quality was key to the progression of the project and was a step change in understanding for both the proposed supplier and my organisation. Key outputs were defined and measured

At the start of the project finished pellet moisture was the primary measure used to determine quality. This was proven to have no correlation with popped cake quality . Analysis enabled the project team to quickly move on from this and re focus efforts.

Illustrating The ProcessD M A I C S

By reviewing the process map, communication with colleagues in the USA and analysis of the process data the critical control points of the process were identified and actions agreed and implemented that is driving both improved understanding of the process and control

26

As we defined the processes critical control points and LT implemented process improvements over time we could see the variability of the finished pellet moisture was reducing . The Low Gel product was also seen to give more consistent running at LT with less extruder blockages as well as improved product after the Skelmersdale popping process.From the point of the failed 22T trial to the success of the September trials there is a statistical difference in the finished pellet moisture showing the greater control and consistency of product now being achieved The Capability graph shows this improvement also. The Pp had

moved form 0.38 to 0.68 and the Ppk had improved to 0.59 from 0.19. This meant the defect level had moved from 22.35% to 2.45%. The data was non normal which LT had explained was due to their scrapping policy of out of specification product.

D M A I C S

Controlling the Improved Process

27

D M A I C S

• In the control phase the key thing was to establish robust specifications that would ensure consistently of product that would match successful trial material.

• A bulk density measure was added to the specification as this was seen as a more consistent and reliable method than having separate weight and diameter checks which are difficult and time consuming to complete.

• It was also an opportunity to review the specification and work with the supplier to reduce un necessary restrictions on the process and increase process efficiencies by doing so.

• The process data of LT pellets has been agreed to be captured by tote during production and this will be analysed and reviewed to ensure consistency as we switch to “normal” running.

• Physical popping checks by LT on their popping machines have been incorporated into quality standards. This will ensure that at the point of manufacture any issues with expansion and quality will be seen at the supplier who can root cause this and prevent poor product at the popping process.

28

Bulk Density was seen to be variable between bags which was in turn leading to variable popped cakes weights at the popping process. The popped cake weights were lower than those seen with the existing supplier. The bulk density specification was set to both reduce the variability seen and bring popped cakes weights in line with the current process.

I also successfully changed the finished specification moisture to achieve a number of benefits.As finished pellet moisture was proven not to be critical to popped product quality a reduction in the target moisture and an increased specification limits would a) Improve yields – moving from a 12%

aim to 11% aim saves £26000 a yearb) Widening the specification by 0.5%

would remove an unnecessary restriction on the LT process which was at 5% waste trying to achieve this measure

Key to good product and included in the specification was the testing of the degree of gelatinisation of extruded pellets through KOH testing at LT (See above) . This was set at 2 - 4 and in combination with the popping test at LT gives a robust visual check of quality prior to shipping to the UK

D M A I C S

Overall Approach

Practical Problem Statistical Problem

Statistical SolutionPractical Solution

y f ( x1 , x2 ,..., xk )

26

28

Six Sigma training has changed my thought processes towards problem solving and has made me much more questioning.

It has given me knowledge and tools to help me define and analyse problems that previously I struggled to quantify.

Provided me with a proven structure that gives me clarity in my approach.

Has helped my to influence the direction of projects and senior managers within the business through clear and unambiguous data presentation.

Helped me to understand process capability indices and how they are applied within the business.

My Experience

The change to six sigma thinking

Old Behaviours

1. We only use experience, not data.

2. We collect data, but just look at the numbers.

New Behaviours

1.We are driven by data in all aspects of our

business.

2.We make decisions based upon facts and not

instinct.

26

Lean 6 Sigma Green Belt

What is 6 Sigma