Multivariate Control Charts in Chemical Batch Processing

Southern Cross

Involved In Discussions
#1
I work in one of the chemical process industries - we make polymers via batch processing. I have been reading about multivariate control charts as a precursor to possibly using them to track our production.

A typical batch goes something like this:
1. Add stuff to a vessel
2. Heat to temperature xxx
3. Hold till ok for next stage
4. Cool to temperature yyy
5. Add more stuff
6. Heat to temperature zzz
7. Hold till ok (testing for viscosity and other data)
8. Cool and add more stuff
9. Adjust to specification and then pack off

I could consider a vast array of data that could be collected, eg. total cycle time, stage 3 hold time, stage 7 hold time, stage 2 heat up time, stage 6 heat up time, data from stage 7, first and final results from stage 9.

Would a T-squared type multivariate chart be suitable for this work? If so, is there any other data that might be added?

Thanks in advance...
 
Elsmar Forum Sponsor
S

sitapaty

#2
I work in one of the chemical process industries - we make polymers via batch processing. I have been reading about multivariate control charts as a precursor to possibly using them to track our production.

A typical batch goes something like this:
1. Add stuff to a vessel
2. Heat to temperature xxx
3. Hold till ok for next stage
4. Cool to temperature yyy
5. Add more stuff
6. Heat to temperature zzz
7. Hold till ok (testing for viscosity and other data)
8. Cool and add more stuff
9. Adjust to specification and then pack off

I could consider a vast array of data that could be collected, eg. total cycle time, stage 3 hold time, stage 7 hold time, stage 2 heat up time, stage 6 heat up time, data from stage 7, first and final results from stage 9.

Would a T-squared type multivariate chart be suitable for this work? If so, is there any other data that might be added?

Thanks in advance...
The data you have to plot are the product characteristics of final output.You are not required to plot the process characeristics.So the type of control chart you need depends on :
a)WHICH IS THE CRITICAL PRODUCT CHARACTERISTIC?
b)Are there more than one critical product characteristic?
On the basis of the process you have described it is difficult to tell which control chart you need.
Sitapaty
 

bobdoering

Stop X-bar/R Madness!!
Trusted Information Resource
#3
I could consider a vast array of data that could be collected, eg. total cycle time, stage 3 hold time, stage 7 hold time, stage 2 heat up time, stage 6 heat up time, data from stage 7, first and final results from stage 9.
I think you are on the right track of considering your variances.

My recommendation is to follow these standards steps to developing SPC:

1. Develop the total variance equation This may include developing a fishbone chart as a tool to help develop the list of variances.
2. Determine which variance factors are adjustable, which are noise, and which can be set as a constant.
3. Minimize the variation of each of the participating variables - get the process in a steady state and capable. Make sure your measurement and gage error is statistically insignificant.
4. Accurately determine the correct distribution of each of the remaining variances
5. Determine which of the remaining variance factors you are going to chart
6. Pick the correct chart to evaluate that variance factor (variable)
 
S

sitapaty

#4
I think you are on the right track of considering your variances.

My recommendation is to follow these standards steps to developing SPC:

1. Develop the total variance equation This may include developing a fishbone chart as a tool to help develop the list of variances.
2. Determine which variance factors are adjustable, which are noise, and which can be set as a constant.
3. Minimize the variation of each of the participating variables - get the process in a steady state and capable. Make sure your measurement and gage error is statistically insignificant.
4. Accurately determine the correct distribution of each of the remaining variances
5. Determine which of the remaining variance factors you are going to chart
6. Pick the correct chart to evaluate that variance factor (variable)
The variances you are talking about are process characteristics (Process inputs)But the control charts deal with product characteristics(Process outputs).SPC control charts tell you whether the process is under control or not.Unfortunately they can not tell you the causes for out of control process.
Sitapaty
 

Southern Cross

Involved In Discussions
#5
Picking out the variables is the easy part. We've been thinking about that for years. I think I need to focus on the mechanics of doing the charts.

A question then: do I separate out the outcome variables from the process variables? In some ways they aren't completely independent. For example, we test for viscosity during stage 7. The cycle time for that stage will depend on how long it takes for viscosity to reach spec. There are process variables during this stage, such as temperature and pressure. The viscosity and cycle times would be more like outcome variables.
 
D

debyang

#6
For a process, you set A’s and measure B’s to understand the process. SPC could be applied on the measured ones, then adjust the settings if necessary. If there were too many to measure, and you have limited resources, then pick those critical to the product specs.
For your example, the cycle time play a part to the viscosity. In such case, you should apply SPC on the viscosity to understand whether the cycle time is appropriate or not.
 

bobdoering

Stop X-bar/R Madness!!
Trusted Information Resource
#7
Sounds to me that a DOE - or at least a CNX evaluation - might help to whittle down the most critical measures. The downside of a t-squared multivari chart is that it will tell you something happened, but not which specific variable caused it. Is that really good enough for you?
 
K

kaikai

#8
Picking out the variables is the easy part. We've been thinking about that for years. I think I need to focus on the mechanics of doing the charts.

A question then: do I separate out the outcome variables from the process variables? In some ways they aren't completely independent. For example, we test for viscosity during stage 7. The cycle time for that stage will depend on how long it takes for viscosity to reach spec. There are process variables during this stage, such as temperature and pressure. The viscosity and cycle times would be more like outcome variables.
Multivariate Control Charts are useful especially when variables are correlated. And of course it can be used to the process variables.

If something happened, you should make scatter plot matrix. It will help you understand what happened in the process.
 
Thread starter Similar threads Forum Replies Date
R Multivariate Control Charts? Control of 2 or more correlated variables Statistical Analysis Tools, Techniques and SPC 6
S Free or Opensource Software that can do Multivariate Capability Analysis Statistical Analysis Tools, Techniques and SPC 8
M How to count the UCL for MEWMA (Multivariate Exponentially Weighted Moving Average) Statistical Analysis Tools, Techniques and SPC 3
H Control of Change i.e. process and WI IATF 16949 - Automotive Quality Systems Standard 9
Sonja D AIAG VDA PFMEA and Control Plan training FMEA and Control Plans 6
F Quality Control in Electrical Calibration ISO 17025 related Discussions 0
Z Why Control Limits are not the same depending on type of exclusion of data points Using Minitab Software 7
D Engineering Change order vs. Change Control ISO 13485:2016 - Medical Device Quality Management Systems 3
D Computer access, password control ISO 13485:2016 - Medical Device Quality Management Systems 6
A How to control distribution list? Document Control Systems, Procedures, Forms and Templates 2
A Control plan for IATF - example? FMEA and Control Plans 1
J ISO 9001:2015 Clause 8.5.1 Control of Production and service provision - Help with Work Instruction Access ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 13
A Final quality control corrective actions ISO 9000, ISO 9001, and ISO 9004 Quality Management Systems Standards 10
J Training on the importance of Document Control Document Control Systems, Procedures, Forms and Templates 3
G How to implement H&S and Quality Control Requirements in Contract for Potential Supplier? Contract Review Process 6
D GAMP CAT 5 Change Control Qualification and Validation (including 21 CFR Part 11) 9
D Use Error Risk Controls and Control Verification ISO 14971 - Medical Device Risk Management 6
S Maintenence related action in Control plan FMEA and Control Plans 3
M ISO 13485 and document control for programs ISO 13485:2016 - Medical Device Quality Management Systems 6
R Control Methods in Control Plan FMEA and Control Plans 5
M Design Control - Management of Electronic CAD files Design and Development of Products and Processes 14
S TSCA (Toxic Substances Control Act) PIP 3:1 RoHS, REACH, ELV, IMDS and Restricted Substances 0
S Use of Attributes Control Charts Statistical Analysis Tools, Techniques and SPC 1
U Appraiser Out of Control Gage R&R (GR&R) and MSA (Measurement Systems Analysis) 5
L Looking for Control Plan and FMEA Stamping process FMEA and Control Plans 2
Z AS9100 Control of Nonconforming Outputs - Rework Dispositions AS9100, IAQG, NADCAP and Aerospace related Standards and Requirements 8
K ASME, ANSI, ASTM and similar specifications and their requirements for document control. Inspection, Prints (Drawings), Testing, Sampling and Related Topics 8
Z Software for design control ISO 13485:2016 - Medical Device Quality Management Systems 5
K Quality Control/Assurance in the Metal Industry Quality Tools, Improvement and Analysis 10
K Defining risk control measures IEC 62304 - Medical Device Software Life Cycle Processes 14
J Warehouse temperature and humidity control EU Medical Device Regulations 7
Jimmy123 IATF16949 Control Plan FMEA and Control Plans 1
D Control Number for Class III and IV devices Canada Medical Device Regulations 1
J Document Control Metrics Document Control Systems, Procedures, Forms and Templates 3
I Document Control Software Document Control Systems, Procedures, Forms and Templates 2
T Controlling Expandable Forms in Paper-Based Document Control System Document Control Systems, Procedures, Forms and Templates 10
N ISO 13485 7.3.9 Change control in medical device software ISO 13485:2016 - Medical Device Quality Management Systems 6
C UDI Questions on Control Units Other US Medical Device Regulations 6
A Quality Control Datasheets Inspection, Prints (Drawings), Testing, Sampling and Related Topics 2
M Control-self assessment Internal Auditing 5
GreatNate Master Control QMS software Quality Tools, Improvement and Analysis 1
A Prototype control plan FMEA and Control Plans 2
J Records Control - Does each individual record need to be numbered? Records and Data - Quality, Legal and Other Evidence 2
shimonv Document Control Procedure Header Content Document Control Systems, Procedures, Forms and Templates 4
W FMEA - Current control and occurrence rating FMEA and Control Plans 3
S Which department should prepare the control plan? could you show me a standard regarding to this matter. FMEA and Control Plans 17
G Control Plan & PFMEA Review Procedure? FMEA and Control Plans 8
N What are the software audit and control steps Reliability Analysis - Predictions, Testing and Standards 2
A Fabric roll inspection - What type of Control Chart to use? Inspection, Prints (Drawings), Testing, Sampling and Related Topics 2
C Monitoring and Control Instruments RoHS, REACH, ELV, IMDS and Restricted Substances 3

Similar threads

Top Bottom