Article:

Implementing Statistical Process Control

Implementing Statistical Process Control linked with Continuous Improvement (SPC-CI) in the food processing plant need not be a complex process. However, if the maximum effect is to be achieved and sustained, SPC-CI must be implemented in a systematic manner with the active involvement of all employees from line associates to executive management.

The ultimate objectives of the SPC-CI process are the following:

Process and product data are reported as a mean and standard deviation
Manufacturing process variation is stable or predictable.
The manufacturing process is capable of producing product that is in specification. This is demonstrated using histograms and capability indices.
Control charts or graphical trend analysis are used to understand changes in the process mean and process standard deviation.
Line associates at multiple locations on the plant floor use control charts. The charts are used to determine when to adjust the process. This is done by determining if the process mean or standard deviation is trending in a way that may cause product deviation. In highly capable processes, it is possible to take corrective actions prior to manufacturing a product that is out of specification. In addition, both line associates and management used the control charts to identify opportunities for process improvement. Management uses control charts to ensure that the SPC-CI process is properly functioning.
Process variation is seen as the problem. It adds needless cost to the product and is the reason why a product does not comply with specifications.
A systematic procedure that focuses on reducing variation is used to continuously improve manufacturing processes.

The objective of SPC-CI differs significantly from the traditional QC/QA process. In the traditional process, the QC/QA tests are used to generate a datum point and this datum point is compared to a standard. If the point is out of specification, then action is taken on the product and action may be taken on the process.

To move from the traditional QC/QA process to SPC-CI, a plan should be developed implemented and followed. This plan should address the following:

Leadership
Expert knowledge
Education and training
Software and hardware
Implementation plan
Process control plan
Systematic process improvement
Financial aspects of poor quality

Leadership

Leadership is critical for the successful implementation of SPC-CI. One of the basic objectives is to eliminate the idea that SPC-CI is a "flavor of the month" and "this to shall pass." Leadership responsibilities include the following:

Walking the talk with regard to continuous improvement - Managers need to show commitment to continuous improvement by taking actions to improve their personal management processes.
Insisting on fact based reporting of data. Data should be reported as a mean and standard deviation. Trends should be reported on control charts
Allocating resources to SPC-CI and removing roadblocks that inhibit implementation.
Setting priorities for continuous improvement. This can be done through a quality council.
Conducting management reviews on the continuous improvement process. This may be done through quality councils
Establishing a measurement system to assess SPC-CI implementation.
Appointing continuous improvement teams, monitoring activities, and ensuring celebrations for successful completion of projects.
Gaining experience in continuous improvement by serving as a member on a continuous improvement team.
Ensuring inclusion of external suppliers (as appropriate) on continuous improvement teams. Ensure that internal customers and suppliers are part of continuous improvement teams
Showing commitment to continuous improvement at the plant level by being present at celebrations.
Developing and managing systems to ensure the communication of continuous improvement success stories to other plants.
Ensuring alignment of the rewards and recognition systems to process control, process capability and continuous improvement achievements.
Conducting an annual audit of the SPC-CI system.

Expert Knowledge

Companies implementing SPC, should have access to expert knowledge in the areas of process control and statistics. This can be achieved by either utilizing internal consultants or external consultants. SPC is based on statistical theory, process control theory and the proper application of the theory to the manufacturing environment. The proper application of these theories should be started at the beginning of the process so that bad habits do not develop and deride later efforts.

The author has observed numerous examples of either poor knowledge of the theory or poor application of the theory. These problems are not usually fatal at first, however, they can lead to serious issues as more complex problems are solved. When the fatal incident occurs, a typical response is that "SPC does not work."

The SPC expert should also select and approve texts for the company since many SPC texts have statistical errors.

Education and Training

Education and training in SPC-CI is a critical factor to the success of the implementation of the process and to maintaining the SPC-CI system. The company should be willing to make a long-term commitment to SPC-CI education. This is needed to deal with turnover of personnel. In addition, the education requirements are depended on the responsibility of the employee. The following training objectives are for the basic skill set needed by associates using SPC on the processing line and for SPC-CI leaders. The SPC-CI leaders are team leaders, who will lead teams to improve the process.

Basic skill set
- Basic math
- Understanding central tendency (mean, median, mode) and variation (range, standard deviation)
- Understanding trends
- Describing the process
- Describing sources of variation
- Interpreting a control chart
- Interpreting a histogram and capability indices
- Interpreting a Pareto diagram
- Functioning in a continuous improvement team
Skill set for SPC-CI leaders
- Basic
  - Basic math
  - Understanding central tendency (mean, median, mode) and variation (range, standard deviation)
  - Understanding trends
  - Understand statistical thinking
  - Process mapping and understand the process
    - Process analysis
    - Draw a detailed flow diagram
  - Cause and effect diagrams
    - Understanding the sources of variation
    - Drawing a cause and effect diagram
  - Control charts
    - Understanding the types of control charts
    - Understanding of the proper selection of control charts
    - Understanding if a process is stable
    - Understanding how to interpret a control chart
    - Understanding how to calculate and interpret a color (rainbow chart) at locations that use the chart
  - Histograms and capability analysis (Cp and Cpk)
    - Understanding how to interpret a histogram
    - Understanding how to interpret capability indices
    - Understanding if a process is capable of meeting specifications
  - Collecting of data
  - Analysis of data utilizing the following techniques:
    - Pareto diagram
    - Scatter diagrams
    - Check sheets
  - Systematic problem solving
    - Planning a continuous improvement project
    - Selecting a cross functional team
    - Using project management techniques
    - Running efficient meetings
    - Utilizing the PDCA Cycle
    - Conducting problem analysis
    - Conducting gap analysis
    - Identifying root cause(s)
    - Formulating several potential solutions
    - Selecting best solution
    - Testing
    - Standardizing results
  - Understanding how to mine data
  - Developing a process control plan
- Intermediate
  - Understanding the use SPC to control manufacturing processes
  - Understanding advanced control chart theory
    - Recalculation of control chart limits
    - Rational subsample
  - Understanding measurement systems
  - Understanding sampling systems
  - Utilizing quality function deployment
  - Utilizing failure mode and effect analysis
- Advanced
  - Using design of experiments
    - Processes must be stable before DOE can be effectively done in production

Software and Hardware

Companies usually face a dilemma in implementing SPC-CI. A decision must be made as to whether to plot the data on control charts by hand using graph paper, or plot the data using SPC software.

Some companies elect to delay installing SPC software to "save" money. A SPC software and hardware system can cost around $1200 to $1400 per site, while a piece of SPC graph paper costs around $1.00.

However, when the decision is made to use a paper charting process, the company does not account for a large number of hidden costs which include:

The time it takes to plot the data.
The time it takes to interpret the chart (SPC software automatically interprets the data.
The time it takes to determine a root cause of a problem when the data are not plotted correctly.
The time it takes to enter the data manually into a computer to generate a histogram and calculate process capability indices.
The time it takes to generate additional data analysis graphs for the continuous improvement process.

Thus, the initial short-term savings that was achieved by delaying the purchase of software and hardware may cause medium term losses.

There is another reason for the immediate implementation of SPC software and hardware. An old saying is that the easy things get done first. Plotting control charts by hand is not an easy task. It takes time and concentration. Thus, if a company makes this process more difficult by hand charting, this may be one more reason for line associates not to comply with the process of SPC implementation.

There are a number of factors in selecting SPC software. Several critical factors include software support, flexibility of integrating data collection systems including automatic data collection systems, and ease of learning the software. If the software is not easy to learn, this will cause needless increase in the training costs.

Implementation Plan

Time must be allocated to properly implement SPC-CI. It takes time to learn the new skills, and time to properly apply the skills. The author has observed, that repeated training classes might be necessary to ensure that the employees properly learn to apply SPC-CI techniques. Old habits must be replaced with new habits.

In implementing SPC-CI in a multiple plant scenario, it is recommended not to try and implement SPC at all plants at once. This recommendation is made for several reasons. Managers must learn new skills in managing facilities using SPC. This must be done both at the plant level and the corporate level. Learning and applying these new skills must be done in addition to the current work responsibilities. Having too large of an implementation process will result in over extending the managers, thus potentially leading to failure.

Using a phased-in implementation system, allows a systematic and sustained implementation plan. The initial plants selected for implementation allow the company identify the lessons learned, the lessons failed, and improve the SPC-CI implementation process. In addition, it allows the company to build success stories, which can be used to further motivate the additional plants to implement SPC-CI.

As part of the implementation process, the resources that are needed to properly implement SPC-CI include:

Selecting hardware and software
Educating and training of line associates
Educating and training of management
Developing a process control plan
Utilizing control charts to make adjustments on the process
Utilizing control charts, histograms and capability indices to identify areas for improvement
Selecting continuous improvement projects
Gathering the success stories of several cycles of continuous improvement
Improving the implementation process
Communicating the success to other facilities in the company - It is recommended that companies utilize the line associates involved in the continuous improvement projects to communicate the success stories.

Process Control Plan

The process control plan is the heart of the SPC process. It defines the process measurements used to assess the process or the product, and actions taken if the measurement signals that a change has taken place in the process.

Systematic Process Improvement

A systematic process should be used to drive the continuous improvement process. This should be done from the very start of the implementation process. Early projects can be defined as picking the low hanging fruit, or projects where it is easy to have success. Associates may question why do they have to follow the "continuous improvement drill." The reason why they need to follow the drill is to develop the skill set and discipline to solve more complex problems. The following is an eight-step continuous improvement drill:

Define the problem in context of the process and the system.
Select a process improvement team.
Document the process.
Measure product and process performance
Understand why the problem exists
Develop hypotheses and test the best hypothesis
Evaluate test results; implement and standardize the best solution.
Celebrate the success and start all over again.

Financial Aspects of Poor Quality

Typically, implementing a "cost of poor quality program" is not needed at first. It is easy to develop a rough estimate of the cost of poor quality for most of the initial problems that must be solved. Usually these rough estimates will actually underestimate the actual cost of poor quality. As the quality improvement process matures, it is recommended that the company start to systematically estimate the cost of poor quality. These estimates can be used as financial drivers for to achieve larger gains.