1. Statistical Techniques
TM-4000-S209
Statistical Techniques

2. First Law of Statistics
Statistics Lie!
Is the sample size tested or inspected representative of daily production?
Is the data that is being collected relevant to your processes?
Collected data is not enough!
What is the data telling you?

3. Quality…How Do You Know?
Verbal reports: QC department, managers, line supervisors and plant personnel…but anectodal is not enough!
Empirical evidence: Written reports, Inspection Forms, Statistical Analysis (variable control charts, spreadsheets, Pareto charts), Customer complaints.

4. Quality is Everyone’s Responsibility!
Quality Control Department, Line Managers, Supervisors, Plant Personnel. Everyone can collect data!
Raw Material Receiving
In-Process Evaluations
Final Product

5. Purpose
To ensure consistent quality and appropriate process control through the use of statistical techniques.
To collect, analyze and interpret data relating to product and process characteristics.
To identify problem areas to reduce returns and warranty claims.
Assist in conducting root cause analysis for problem areas.

6. Purpose
The general approach to on-line quality control is straightforward: simply extract samples of a certain size from the ongoing receiving, production, final inspection, shipping, in-field performance and customer feedback processes. Line or variable control charts of the variability in those samples will identify product specification compliance, and consider their closeness to target specifications. If a trend emerges in those lines, or if samples fall outside pre-specified limits, then the process is deemed to be out of control and action is required to find the cause of the problem.

7. Tools
Variable control charts, spreadsheets, visual observations derived from data gathered from processes, testing and inspection.
Identification of special causes of variation or deviation from the required specification.
Identify sources of product variation and group into two major causes: common and special causes.

8. Test of Inspection Name
Responsibility
Data Type
Examples
Raw material receiving
QC or designated personnel
Test values
Visual
Components such as connectors, spacers, desiccant, sealants etc.
Damaged containers
In-Process evaluations
Water temperature, Cutting Wheels, Component Testing (ie sealant adhesion)
Final product
Skips and gaps in sealant
Field Performance
After Sales Service and QC
Written Inspection Reports
Verbal reports
Seal failure, volatile fogging
Customer Complaints
Verbal and electronic correspondence
Delivery delays, seal failures, volatile fogging, after sales service

9. Raw Material Receiving
Maintain Raw Material Log: date received, component received, manufacturer identification, lot tracking numbers (if applicable), condition of packaging material, condition of material – conforming or nonconforming to specifications
Variable Control Charts: used to evaluate variation in a product where the measurement is a variable--i.e. the variable can be measured on a continuous scale (e.g. height, weight, length, concentration).

10. Variable Control Charts
The general principle for establishing control limits applies to all control charts. After deciding on the characteristic you want to control, for example, the standard deviation (acceptable upper and lower limits within the specification), you estimate the expected variability of the respective characteristic in samples of the size. Those estimates are then used to establish the control limits on the chart. *
*Center line indicates product conforms to required specifications. UCL and LCL indicate upper and lower data points for acceptable product.

11. Raw Material Receiving
Sampling: Determine how much needs to be tested which will be representative of the entire lot or batch.
Nonconforming Product: Identify nonconforming product, is the nonconforming product within tolerance or acceptable levels?, isolate it so it will not be used if outside of accepted specification, contact manufacturer of product.
Isolated or Repetitive Issue?: Variable control chart will indicate if nonconforming product is an ongoing or isolated issue.

12. In-process Inspections
Designated plant personnel: collect data and complete forms such as seal adhesion, desiccant capacity, was spacer stored properly, cut glass sizes, gas content verification, etc.
Variable Control Charts: monitor the extent to which the components that are used to fabricate your products meet specifications.

13. In-process Inspections
Quality Control Forms: The following information is required for each insulating glass unit tested:
Length
Width
Position of spacer
Glass thickness
Airspace
Glass Edges
Primary and secondary sealant
Glass coatings
Percentage gas fill required and achieved

14. Final Product Inspection
Data is collected from an appropriate sample size based on your production; the suggested number of finished units to be inspected shall be randomly selected as determined from the following:
  Production Number Inspected
up to 25 1
over

15. Final Product Inspection
Each day representative samples shall be inspected for workmanship for at least the following characteristics:
Overall unit size and thickness.
Alignment of glass lites.
Cleanliness of airspace.
Sealant bond to glass and to itself at corners.
Sealant minimum vapor transmission path.
Spacer position (sight line) relative to the unit edge.
Uniformity of sealant application.
Possibility of holes or underfills.
Overall workmanship in finished units or windows.

16. Final Product Inspection
Data is collected on a random basis: representative samples are selected which will represent the daily production.
Pareto Charts: are useful for non-numeric data, such as "cause", "type", or "classification". This tool helps to prioritize where action and process changes should be focused. If one is trying to take action based upon causes of accidents or events, it is generally most helpful to focus efforts on the most frequent causes. Going after an "easy" yet infrequent cause will probably not reap benefits.

17. Pareto Charts
A Pareto Chart is generally shown as a vertical bar chart. A Pareto Chart is a special form of a histogram where the categories have been sorted from most frequent to least frequent. One would not want to sort the categories from most frequent to least frequent if there is a natural order to the categories, such as a distribution by age or cycle time.

18. Trend Exists on Control Chart?
Pareto Charts
Trend Exists on Control Chart?
Data Points Used
Purpose No
Use all data in the applicable or relevant statistical time period
To find common causes(s) to apply to process improvement
Yes
Use only the data for the Point(s) which have been identified as within the significant trend, such as a point outside the control limits.
To find special cause(s) for the determination of corrective actions for trends of decreased compliance or identify successful corrective actions which have resulted in increased product compliance.

19. Final Inspection
Data is prioritized based on company quality control requirements and specifications.
Data is reviewed by all! (management, line supervisors and plant personnel)
Data which identifies that the level of product rejection is above the established, acceptable levels in sent to facility management for review, discussion and action.

20. Conclusion
It’s not enough to collect data! Data is meaningless unless it is reviewed, analyzed (what is it telling you) and acted upon.
Collect the right data and the right amount of data! One data point does not provide you with enough information to determine if a component or product conforms to specifications.
Don’t assume you know the cause of a problem!
Statistical analysis can help you control nonconforming components and products.