To reach Six Sigma decisions, consideration must be given to the complete product manufacturing process and how proposed change impacts the total process.

In today's world where product design information is entirely

digital, it seems incomprehensible that the total product requirements-engineering definition plus manufacturing and quality assessment processes-are not digitally associated. Yet, in most cases, design and manufacturing personnel still work in silos with only a loose IT connection. There is often little or no connectivity between computer-based applications to enable truly seamless collaboration. Instead, many companies have complex and labor-intensive business processes that require manual transfer and reconciliation of information between departments and applications. Tasks are compartmentalized by department and performed serially.

Worse yet is that data is not transferred and the correct or most current information is not provided to the next operation in a serial process methodology. Information is either lost, misplaced, perceived as not important to the next department or simply does not pass the firewalls set in place by business practices, job security or personal protection.

This current scenario adds unnecessary elements to the process, thereby increasing the risk of error. With this setup, a manufacturer has:

-Multiple people handling data

-Multiple departments handling data

-Manual transfer of data

-Reconciliation of data

-Validation of data transfer

Problems abound in such environments. The need of operators in multiple departments to manually transfer and reconcile data is error prone, labor intensive, costly and slow. As many as 60% of all quality errors are because of a failure to accurately communicate engineering requirements during new product introduction or during design changes. Errors often go undetected until the product is in production, resulting in costly changes on the shop floor. Even worse, errors go undetected in the form of quality "escapes" that are significantly more costly to correct.



One Becomes Many

Even a simple component can require hundreds of characteristics to complete a total engineering definition. The need to produce a hole in a specific part is clearly specified and is considered a single feature as well as the computer-aided design annotation object associated with the hole. From a product definition perspective, the hole is considered a single aspect of the product design. However, in the manufacturing view there are at least five distinct characteristics to describe the hole. All five will require supporting process and quality planning to account for each individual characteristic.

The current process for defining and changing characteristics is labor intensive and slow, typically requiring six to 10 hours to define the characteristics on a new drawing and three to four hours to work through a change in design. Additionally, the requirements that originate from specifications and standards makes comprehending the quantity of characteristics an overwhelming task. Only portions of each document are applicable to the specific part or manufacturing process. Documents often contain conflicting requirements, which can be ambiguous, complex and subject to interpretation. It takes time to sort through the separate specification documents and missing the characteristic details has significant impact on quality.

To accurately identify the characteristic requirements, the manufacturing engineer typically uses a manual process called ballooning. This process allows the manufacturing engineer to identify and tag each characteristic on the drawing by creating a separate identification label-or balloon-for each characteristic. The entire process can take weeks to complete. And, if a single feature is missed during this tedious process, then quality is degraded.



A Quality Plan is Not Enough

Some potential pitfalls can be avoided if a good Quality Plan is in place. A Quality Plan is a document developed to identify and provide assurance of conformance of all engineering characteristics. The Quality Plan must be maintained over the life of the component. The process must be repeated when there is a change in design or significant change in the manufacturing process.

The current process requires about 60 hours to create a Quality Plan for a new product and 30 hours to revise one for a typical design change. Additionally, the current quality planning process is error-prone-two to three sigma-with a degree of variability that causes missed requirements as well as incorrect and inconsistent interpretations of key characteristics. If an error is made at any step in the process, the error is compounded through the supply chain.

Several types of critical errors can occur during the quality planning process:

-Missed characteristics

-Incorrect interpretation of characteristics, most commonly in specifications

-Poor change incorporation

-Inadequate gaging or verification methods

-Unclear or confusing work instructions



In the Cost of Quality Pyramid, it becomes evident that the higher an escape penetrates into the supply chain, the greater the financial impact. These costs depict only cost of correction to the issue causing the quality escape. An additional cost of quality escape is associated with compensation for lost revenue in transportation fleets or liability costs associated with more catastrophic events.

For example, the amount of time required to create new or revise existing work instructions falls into three categories based on complexity of the component or operation: simple, medium complexity or complex. A typical aerospace company, by conservative estimate, creates approximately 8,000 to 10,000 work-instruction documents and incorporates 25,000 to 35,000 revisions during the year.

These work instructions are created in silos using information handed off manually from other departments. Like other steps in the future state, the process is labor intensive and error prone. Undetected errors and omissions in work instructions frequently result in quality problems and rework in production. Each of those changes must be correctly incorporated into the supply chain definitions such as the Quality Plan and work instructions. For industry at large, no solid data is available, but it is safe to say that managing escape occurrences has not reached the Six-Sigma level.

For an idea of how much potential for error exists, a typical automotive engine manufacturer manages:

-50,000 active part numbers

-30,000 parts by suppliers

-500 characteristics per part

-26,000,000 active characteristics

-6,000 design changes

-5,000 tooling changes

-8,000 numerical control

program changes

-3,000 source changes



Overall Cost Benefits

Based on the Value Drivers in table, "Value Drivers of Digital Manufacturing Technologies," the total quality and

productivity-related savings opportunity for an enterprise managing 10,000 part numbers is estimated between $11.1 and $15.6 million per year. Even manufacturers of a single product can experience variations within models that produce challenges to normal operations, where multiple part numbers must be maintained with individual characteristics for each part.

For an example of what this means from a cost standpoint, a defense OEM might manage 10,000 part numbers with an average characteristic count of 500 for a total of 5 million opportunities for a characteristic escape. Using traditional Three-Sigma quality levels, 5 million opportunities will result in 6,750 escape incidents. Cost per incident will vary widely, but if a conservative average of $5,000 is assumed, the total cost is $33.8 million. Extend that across all defense industries and the total estimate annual cost runs into the billions of dollars.

Now assume the same situation using product lifecycle management (PLM) Six Sigma technologies, supported and validated by digital manufacturing. PLM is a strategic business approach that allows for the collaborative creation, management and analysis of product data information across the enterprise. All involved parties are now seamlessly sharing the most up-to-date data using consistent Six Sigma business practices and problem-solving methodologies.

This company now achieves Six-Sigma quality levels to manage characteristics. Quality escapes from characteristic mismanagement will reduce to 17 incidents and if the average cost per incident remains $5,000, total cost will be $85,000 and savings will be $33.7 million. It is clear why PLM Six-Sigma, along with digital manufacturing, bring immense value propositions. The potential benefit to other participants in the product lifecycle including OEM suppliers is enormous.



High Cost of Low Quality

In 2003, there were 19.5 million recalled vehicles with a total of 529 vehicle types affected. Automakers face a staggering $12 billion annual bill to fix vehicles covered by warranty and there is growing tension with OEM suppliers as to who is responsible for the skyrocketing cost and how to reduce it.

While not all defects in parts can be identified or corrected by digital solutions, the ones that can be are the most cost effective. The concept of passing the warranty cost where the fault occurs can be neutralized by the elimination of the cost by reducing and ultimately, eliminating the defects.

In manufacturing any product, assembly production problems account for more than 30% of all warranty costs because of incorrect sequence, incorrect part or incorrect procedures. Typically, these assembly problems are underreported as much as 20% where they are rigged in process without notification to maintain or just due to lack of visibility where the error escapes go unnoticed for later discovery.

Topping any corrective warranty cost list from an industry analyst, consultant or Six Sigma company truly implementing the available technologies are:

-Preventive engineering activities that are a collaborative process between manufacturing and product development.

-Rapid problem elimination-the faster a manufacturing or design problem is identified, the faster it is contained and corrected with automatic change propagation, and the smaller the contamination.



Impacting Relationships

Regardless of the methodologies used to support a Six Sigma process, they all rely on a collection of data, statistics, analyses, charts and numbers. While these metrics can be reviewed and compared, it is a challenge to comprehend the effect that one change can have on all disciplines in the process. Based on numbers alone, it is nearly impossible for anyone to see how a single change he recommends in his area impacts the supporting processes.

To reach Six Sigma decisions, consideration must be given to the complete product manufacturing process and how any proposed change impacts the total process. Visualization is the common denominator among all disciplines of an enterprise. Three-dimensional graphic representations of the Six Sigma methodology's data provide a clear and accurate accounting of proposed or active changes for improvements in a manufacturing process. Applying digital solutions is not a waste of time for lean manufacturing or in support of Six Sigma quality. Q



Quality Tech Tips

-As many as 60% of all quality errors are because of a failure to accurately communicate engineering requirements during new product introduction or during design changes.

-Errors often go undetected until the product is in production, resulting in costly changes on the shop floor.

-Product lifecycle management allows all parties to seamlessly share the most up-to-date data using consistent Six Sigma business practices and problem-solving methodologies.

-While not all defects in parts can be identified or corrected by digital solutions, the ones that can be are the most cost effective.



Value Drivers of Digital Manufacturing Technologies

-Instantaneous notification when design changes are completed at the product design level, signaling manufacturing to initiate changes impacting manufacturing processes.

-Ability to view current drawings as soon as design changes are completed to establish requirements for processes, tooling and fixturing.

-Ability to extract characteristic data from drawings to validate and verify manufacturing and process schemes.

-Faster and less error-prone updates of quality plans, numerical control programs and work instructions.

-Streamline first-article inspection including confirmation of compliance to all characteristics and variants.