Capability Analysis for Packing Process in Food Manufacturing

D

Dave.

#1
Hi, I am new to the forum and looking for some help please. I am currently working at a UK food manufacturing company looking at their packing process and I am wanting to perform a capability analysis. I am relatively new to Six Sigma (but not Lean), and I am a bit stumped.

In the UK, when products are packaged to a nominal weight, manufacturers can use the average packing system, where they must conform to 3 packers rules by law.

1. The average weight of goods must not be less than the nominal weight (Qt).

2. 2.5% of goods can weigh between T1 and T2.

3. No goods can weigh less than T2.

where
T1 = Qt-TNE
T2 = Qt-2*TNE

and
TNE = 1.5% of Qt for products 1000g to 10,000g

I have taken 50 measurements (see end of post) of product weights as they roll off the process and plotted an I-MR chart which shows some special causes. The problem is now that I am struggling to do a Capability analysis of this data. I know that yo need to have a process in control before producing a Capability Analysis, but I thought whilst I was doing the I-MR chart, I would get 100% familiar with CA as it will be useful in other food companies I am working at.

The problem is that whilst the measurements taken follow a normal distribution, I don't want them to and the reason is this.

Say the target weight =2500g, T1=2462.5g and T2=2425

Therefore I am taking my LSL to be 2425g as this is the minimum legal requirement. As this is a food product there is no USL from the customer as this is Giveaway. The manufacturer however wants to limit Giveaway to 0.8%. This gives me a USL of 2520g.

Technically speaking, for this process to be in control, it needs to be Skewed to the right, with the peak just above the target, within the USL and LSL, with 2.5% of the left tail between T1 and T2, and nothing below T2. Cp, Cpk, Pp, Ppk and Cpm will not be a fair representation of this process if a Normal Distribution is applied.

I presume there is a way of doing this as this is a common theme in many food factories that pack to the average weight system, but after 3 days of searching on line I haven't really come up with anything.

Please can some one point me in the right direction and explain. I did post on another forum and all I got was "try looking at Cpm".....Not much online about that specifically....either that or I am just looking in the wrong places.

Thanks in advance for your time and guidance.

Dave.


Measurements:
2504.2 2512.6 2465.1 2462.9 2450.9 2501.6 2473.7 2470.1 2463.2 2474.9 2471.2 2483.8 2489.4 2463.5 2500.4 2499.5 2530.1 2495.8 2492 2500.6 2533 2526.1 2510 2495 2506.9 2507.9 2492.4 2488.3 2494.7 2512.3 2496 2527.8 2537.8 2493.3 2490.2 2459.1 2485.1 2483.7 2442.7 2494.9
 
Elsmar Forum Sponsor
G

George Weiss

#2
Re: Capability Analysis in Food Manufacturing HELP!

Greeting Dave,[first-post person],
You have found a place for answers, but they are already sleeping in the USA. Searching the Eslmar Cove is a link below @
Process Capability Study using Variables Data (weight of the product)
You will also learn by finding this link your self.
Press the [Search] button, and ask for "capability study".
I am sure the experts will be answering this thread in the morning Friday, or sooner..............
I am guessing you have already perfomed an MSA study, comment @
How to use Sampling Plans - Confusion on Lot Sampling
I hope this helps............
 
Last edited by a moderator:
D

Dave.

#3
Re: Capability Analysis in Food Manufacturing HELP!

Thanks for the reply, helpful but still not giving me an answer? As for MSA....thats a good question, No I haven't as I did not consider using SPC until I had left the company, however that is something that I will do when I return. This however raises another question:

The measurement system does not require "human" interaction other than when the process is set up, eg. Products roll down the line and if they are too light they are rejected and if they are too heavy they are rejected. Now the limits that are being used (I think) do not conform to the packers rules or the USL of 0.08% Giveaway, this is something that I need to investigate.

The problem that we are having is that the throughput of the line is being crippled by the fact that the product weights vary so much creating rework (minimising Giveaway). When this happens, the line leader presses a button to accept overweights reducing rejects, which increases throughput but crippling Giveaway.

Having not done and MSA, I presumed (probably incorrectly) that the measurement system must be OK as the weights are a legal requirement....however after reading some posts on here I think I will do one when I am next there.

The line leader suggests that the quality of the product coming to the line is older than it should be due to the fact the the previous process produced it too early and thus it having to wait in a WIP area. One would suggest then to solve this problem level the production between the 2 processes so that mixing supply the packing operation with only fresh product as and when it is reqiured. However, there are a multitude of things that can affect the consistency of the product when it is mixed, some that could be controlled better and some that I doubt could. Even if the control of product from the mixing process could be controlled better, I doubt that it will yield perfect results every time, leading me to investigate if the cause of variation is the way in which it is "segmented" when it is to be packed. When I looked at the weight range of product that was new and product that was "old" they were pretty similar +/- 10g which is what leads me to believe that it is not the mix consistency that is the issue.

Bulk product is tipped into a hopper and then extruded. This is then cut into slugs using a timed cut rather than a volumetric flow rate to signal slug cutting. I have been told that the machine really cannot be changed, however, I think that the cutter for starters is in the wrong place. If I can show through SPC that the variation experienced is natural for the packing process, then it will save time and effort trying to solve the unsolvable when a new/better extrusion/cutting system is needed, or the Specification Limits need to be adjusted (specifically the USL where 0.08% Giveaway may be too tight).

Am I on the right lines? As I said, I am new to 6 sigma/SPC in a practical sense (did it at university many moons ago), however the more food companies I go into, the more I think this kind of analysis is needed. Just for your information, my job is as a Lean Engineer making improvements. The companies I work in (some large and some small) have absolutely terrible measuring systems if at all; One company didn't measure Rework as it was not seen as a problem ("...well its sorted straight away....". It turned out to be around 40-50% across all 6 lines and they wondered why they were struggling for capacity!!

Thanks for you help.

Dave.
 
Last edited by a moderator:

bobdoering

Stop X-bar/R Madness!!
Trusted Information Resource
#4
Re: Capability Analysis for packing process in Food Manufacturing - HELP!

This issue may take some digestion - pardon the pun. First of all, I almost got caught up in the trap of jumping right to calculating the capability - ignoring my own best advice. Fact is, the first step is developing the Total Variance Equation - and you are going to need to do that. My stab at starting it is attached. That will help you recognize that the variation that you measure is multi-modal, which factors are the major factors of variation, which ones can or can not be controlled, and that the capability is a function of all of these factors and not just one "output" variation. Weight measurement MSA is the first factor I included.

The variances I come up with on the attached equation would all be normally distributed. Therefore, at this point you would expect the output to be normally distributed. The distribution analysis shows (by p=.9734) that it undoubtedly can be modeled with the normal distribution.

When you mentioned extruding, capability really gets complicated. You add a significant variance with density extruding - especially if it is highly compressible material. That is going to confound "dialing in" weights!

From what I can tell, there are no controls that would limit any of the inputs specifically to your legal requirements, except for setup and sorting. There are no controls that can change your distribution curve except truncating by sorting. So, to get to your goal, you need to determine what new controls you can develop to minimize variation or change your distribution. You have quite a challenge.

For those that want to play along, I provided an excel sheet with the data.
 

Attachments

Last edited:
D

Dave.

#5
Thank you so much for the help, please could you attach the word doc again as I only have 2003 so can't read docx files....it makes me laugh as I am a complete novice when it come to statistics.

This, as I see it is a common issue in most food manufacturers that I have worked in, whether something is extruded, poured or added to make up a product, so I assumed all of the maths and kinks would have been ironed out to a handful of equations to simply plug some numbers in and compare some values.....:lmao: It appears I was wrong.

So if I understand you correctly, I do an MSA on the ingredient weighing process prior to mixing to identify any issues there. Most of the bulk ingredients are fed from silos which is computer controlled, with additives etc weighed out by hand. I assume that the company will have well established standards for what the tolerances should be but maybe the operators are not rigidly sticking to them, or systems not calibrated!

Once that is done and I am happy that any issues are remedied (equipment calibrated, operators "trained" etc), I then do an I-MR on the ingredient weighing process for each ingredient and follow that batch and do an I-MR on the packing process to compare the results, (I take it when you say "variance" you are talking about the standard deviation?).

From this I will be able to identify what it may be at the mixing stage that causes the greatest swings in variation by performing a capability analysis (At the mixing stage as this as you say will be normally distributed data). If the mixing process is "capable" of knocking out what it should then the issue lies somewhere else other than with the way in which the ingredients are mixed and weighed.....maybe the length of time product waits or, as I suspect the USL at the packing stage is too tight.

If the ingredients mixed are in batches....large batches, what is the best way to collect this data so that it is statistically sound....as an aggregate figure, but measure say at least 50 batches, then follow these 50 batches when packed (1000's of products) and take say 50 consecutive measurements for each batch.

So if all this is correct then Ace, I get it now, if not then I think that maybe I should not open up Pandora's box and stick to what I know...Lean.

If this is something that is likely to take an age, I may have to leave it for another company as we tend to undertake a 4-6 day Kaizen intervention spread over as many weeks (don't like doing blitzes). If companies are new to Lean (as most of them are that I deal with), then this stuff will really blow their minds....however I think for my own purposes it is very useful to know.

Once again thank you very much for your help.....I'm getting there....I think!

Dave.
 
D

Dave.

#9
Still, I am baffled regarding the capability of a packing process producing to the average weight system. How would you perform a capability analysis and compare it to something like this rather than normal distribution? Where:
LSL = T2 (nothing can be below this)
USL is acceptable Giveaway of the manufacturer
Qt is the Target weight (the average weight must be >=Qt)
T1 is the standard to account for variation where only 2.5% can fall bellow this.

Sorry about the quality but just knocked it up in paint.

 

bobdoering

Stop X-bar/R Madness!!
Trusted Information Resource
#10
So if I understand you correctly, I do an MSA on the ingredient weighing process prior to mixing to identify any issues there. Most of the bulk ingredients are fed from silos which is computer controlled, with additives etc weighed out by hand. I assume that the company will have well established standards for what the tolerances should be but maybe the operators are not rigidly sticking to them, or systems not calibrated!
Yes, you need to understand what portion of the variation is from measurement error. Right now you overall variation is 145 or about 6%. Is it weighing error? Is it cutting variation? What can you improve?



Once that is done and I am happy that any issues are remedied (equipment calibrated, operators "trained" etc), I then do an I-MR on the ingredient weighing process for each ingredient and follow that batch and do an I-MR on the packing process to compare the results, (I take it when you say "variance" you are talking about the standard deviation?).

From this I will be able to identify what it may be at the mixing stage that causes the greatest swings in variation by performing a capability analysis (At the mixing stage as this as you say will be normally distributed data). If the mixing process is "capable" of knocking out what it should then the issue lies somewhere else other than with the way in which the ingredients are mixed and weighed.....maybe the length of time product waits or, as I suspect the USL at the packing stage is too tight.

If the ingredients mixed are in batches....large batches, what is the best way to collect this data so that it is statistically sound....as an aggregate figure, but measure say at least 50 batches, then follow these 50 batches when packed (1000's of products) and take say 50 consecutive measurements for each batch.
Depends. Have to go back to the equation. Where are the variations you can control? You need to sort out your CNX variables. Density (mostly a mixing function, if the extruder has no controls)? Cutting length (trust me, it will be tough to measure!)? It is hard to say from a distance. I would do 30 samples for basic verification of capability studies.


So if all this is correct then Ace, I get it now, if not then I think that maybe I should not open up Pandora's box and stick to what I know...Lean.

If this is something that is likely to take an age, I may have to leave it for another company as we tend to undertake a 4-6 day Kaizen intervention spread over as many weeks (don't like doing blitzes). If companies are new to Lean (as most of them are that I deal with), then this stuff will really blow their minds....however I think for my own purposes it is very useful to know.
Well, didn't say it was easy! I agree - this could blow some minds. But, could make some fascinating improvements in throughput and loss prevention. That could be a tidy sum.
 
Thread starter Similar threads Forum Replies Date
J Process capability analysis in ISO 9001:2015 ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 3
PQ Systems Control Charts & Capability Analysis 101 Using SQCpack Software 0
B Two excellent examples of process capability analysis from Quality Magazine Capability, Accuracy and Stability - Processes, Machines, etc. 5
R Capability analysis - What is going on this chart? Manufacturing and Related Processes 15
A Capability Analysis for Packaging Seal Strength with spec. >0.1 Kgf using Minitab Using Minitab Software 6
J Capability Analysis - Unusual Statistical Distribution of my Proccess Capability, Accuracy and Stability - Processes, Machines, etc. 5
C Minitab Binomial Process Capability Analysis Chart Explanation Using Minitab Software 1
C P value not greater than 0.05 for Capability Analysis Statistical Analysis Tools, Techniques and SPC 3
M Sample size when conducting capability analysis with OK and NOK parts Statistical Analysis Tools, Techniques and SPC 3
R Capability Analysis (using Histogram) for Subgrouped Data Statistical Analysis Tools, Techniques and SPC 16
A Inspection Capability Study (Kappa Analysis / Attribute R&R) as Certification Req't Training - Internal, External, Online and Distance Learning 7
C Process Capability - Attribute Data Analysis (Template) Document Control Systems, Procedures, Forms and Templates 0
C Process Capability - Variables Data Analysis (Template) Document Control Systems, Procedures, Forms and Templates 0
A Cnpk (Non-Parametric Capability Analysis) to assess whether the Process is Stable Capability, Accuracy and Stability - Processes, Machines, etc. 7
Q JMP Measurement Capability Analysis Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 1
P Data Analysis - Johnson Transformation/Weibull Distribution/Capability Analysis Capability, Accuracy and Stability - Processes, Machines, etc. 15
L Capability Analysis - Help with Minitab results Using Minitab Software 15
P Capability Analysis for Limited Resolution (thickness of sheets) Capability, Accuracy and Stability - Processes, Machines, etc. 6
M Capability Analysis and Sample Size Capability, Accuracy and Stability - Processes, Machines, etc. 8
B Capability Analysis of a Repeated Feature Quality Tools, Improvement and Analysis 9
B Capability Analysis of a Small Sample Size (as small as 2 and as large as 30) Statistical Analysis Tools, Techniques and SPC 14
S "Normal" vs. "Between/Within" Capability Analysis Capability, Accuracy and Stability - Processes, Machines, etc. 2
S How to conduct Capability Analysis for this data? Statistical Analysis Tools, Techniques and SPC 12
S Free or Opensource Software that can do Multivariate Capability Analysis Statistical Analysis Tools, Techniques and SPC 8
T Capability Analysis in the Real World Capability, Accuracy and Stability - Processes, Machines, etc. 7
I Machine Capability Analysis in Multi-Dimensional Cases Capability, Accuracy and Stability - Processes, Machines, etc. 2
A Yearly MSA (Measurement System Analysis) & Capability Study - TS 16949 Requirements Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 2
S Capability Analysis of Non-Normal Data Statistical Analysis Tools, Techniques and SPC 23
T Process Capability Analysis Binomial or Normal using Minitab 15 Using Minitab Software 54
P Capability Analysis in Minitab - Subgroup size Misc. Quality Assurance and Business Systems Related Topics 2
J Capability Analysis (Cpk) of torque driver process Capability, Accuracy and Stability - Processes, Machines, etc. 5
E Process Capability Analysis for a Service Company ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 7
H Regarding: Describing Graphical Overview and Capability Analysis for main dept leader Quality Manager and Management Related Issues 1
M Macro that will perform a normality test and capability analysis - Minitab help Using Minitab Software 1
C Does a Chronically Unstable Process need a Capability Analysis? Statistical Analysis Tools, Techniques and SPC 10
F Multi Capability analysis with MINITAB Using Minitab Software 2
A Whats the rationale for a Sample size of 30 for Capability Analysis? Statistical Analysis Tools, Techniques and SPC 5
D Normal vs. When a Between / Within Capability analysis should be used? Statistical Analysis Tools, Techniques and SPC 12
A Capability Analysis - Dealing with non-normal data in Minitab Using Minitab Software 8
E MSA vs. Measurement Capability Analysis - What are the differences? Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 33
J What causes Ppk to be greater than Cpk when a basic capability analysis is run? Capability, Accuracy and Stability - Processes, Machines, etc. 17
A Sample size for Capability Analysis in PPAP APQP and PPAP 2
L Definition Cp Cpk Pp and Ppk Definitions - I'm doing SPC and Capability Analysis Definitions, Acronyms, Abbreviations and Interpretations Listed Alphabetically 7
Nihls Capability of a roughness gauge Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 3
H Calibration / Capability in a PCP FMEA and Control Plans 1
G QSB+ : capability of measurement equipment IATF 16949 - Automotive Quality Systems Standard 15
D Data normality versus capability Capability, Accuracy and Stability - Processes, Machines, etc. 11
T Capability calculation of surface profile Capability, Accuracy and Stability - Processes, Machines, etc. 20
PQ Systems Understanding Capability and Control Charts Series Using SQCpack Software 0
J Process Capability - Determining the process capability of certain equipment Lean in Manufacturing and Service Industries 6

Similar threads

Top Bottom