Articles

Effective Implementation of the Automotive "Core Tools"

By Paul Hardiman, General Manager SMMT

For many years automotive quality management system standards, such as QS-9000, VDA6.1 and more recently the global quality management system technical specification ISO/TS 16949, have mandated the use of core tools (Advanced Product Quality Planning (APQP), Failure Mode and Effect Analysis (FMEA), Measurement System Analysis (MSA), Statistical Process Control (SPC), and Production Part Approval Process (PPAP). While some organizations have embraced these tools and seen the benefits they can offer, many other organizations "go through the motions" by filling in the paperwork to get through an audit, without realizing any real business benefits.

This article briefly reviews each of the core tools, their linkages, and how the core tools, if used correctly in a process-based management system, can lead to continuous improvement, defect prevention and the reduction in variation and waste.

First, let's look at APQP. In the implementation of any new product or process, an organization needs to understand the customer requirements. This will include requirements related to technical specifications and proposed timings. If done correctly APQP can help reach the goals to meet the customer requirements (effectiveness) and the organization requirements (efficiency).

APQP is structured around a Plan, Do, Check, Act approach, as shown below:

The "plan and define" phase is critical to a project's success. In this phase a multidisciplinary team should be established, a project manager appointed, and the project timing plan agreed upon. The complexity of the plan will depend on whether the organization or the customer is responsible for the product design. Once the plan is established; and the goals agreed upon, one of the tasks of the team is to consider the potential ways in which a product and/or process could fail. ISO/TS 16949 mandates the use of FMEA to facilitate this.

Some organizations concentrate too much on filling in the FMEA forms, rather than using FMEA as a tool to highlight and prevent future failures from occurring. Let's look at a simple example. The template below shows an extract of the product FMEA for a toaster:

Historically, organizations would have focused on the Risk Priority Number (RPN), which is calculated by multiplying the rating given for Severity of the potential failure, the rating given for likely Occurrence of the potential failure, and the rating given for Detection of the potential failure (SxOxD). In the Toaster example above, this would have focused improvement action on the spring failure (RPN 256). However, if we think about risk to the customer, the greatest risk is corrosion of the heater element (severity rating 9). The AIAG reference manual, FMEA, 4th edition, tries to de-emphasize the focus on RPN, and suggests other criteria could be used, such as Severity x Occurrence (SO), SOD (a non-arithmetic combination of Severity, Occurrence and Detection) or SD (a non-arithmetic combination of Severity and Detection).

Whatever criteria are used the intent of FMEA is to focus on reducing risk of failure occurring by making design and process changes. By focusing on potential risks before the event (prevention) rather than after the event (detection) this approach has been proven to save organizations significant amounts of money by eliminating failures/product recalls.

Now let's look more at a process FMEA:

The above shows an extract from a process FMEA for the manufacture of the toaster. This focuses on the way that the process could fail to meet the specified requirements. The scoring criteria and evaluation are similar to that explained above for the product FMEA. The process controls stated on the FMEA should provide input into a document called a "control plan".

The control plan is a key document defining all the controls required to ensure that the process and parts manufactured are conforming to the required specifications. The controls in the control plan should then be communicated to the operators by work instructions - for example, the method of undertaking the paint thickness checks.

Let's now look at the measurement systems used to ensure that the process/product characteristics are met. The extract from the toaster process FMEA identifies that a paint thickness gauge is to be used to verify the paint thickness on the toaster case. First an organization needs to ensure that the gauge is calibrated (traceable to national or international standards); this could be done internally, or by an external accredited laboratory. Some organizations then assume that this is all that is needed to be done to give confidence in measured values. However, to truly have confidence in the measured values, we need to consider the variation in the measurement system.

The measurement system is the complete process to obtain the measurement. Variation may be introduced through the work environment (temperature, cleanliness, etc.), the people using the gauge, and the method used in the use the gauge. To analyze a measurement system, typically a gauge Repeatability and Reproducibility (R&R) study is undertaken. This type of study evaluates the variation introduced from the equipment (repeatability, equipment variation (EV) and through the appraisers (reproducibility, appraiser variation (AV)). Typically to undertake a study, 10 parts (in this example painted toaster bodies) and three appraisers (the normal users of the equipment) are needed. Each appraiser measures the parts a minimum of twice, in random order, in the normal work environment. The results are then evaluated, using spreadsheets, or specialist statistical software, such as the example below.

For the study results shown above, the gauge R and R was calculated at 4.43%. Typical acceptance criteria is that the gauge R and R should be less than 10% (this would give confidence that the measurement system is detecting variation in the product, and that the variation in the measurement system is small compared with the normal process variation or the tolerance). If the gauge R and R is above 10%, the organization should evaluate the results and take corrective action, which may be to better train appraisers in the use of the gauge (i.e. to reduce appraiser variation AV), or focus on how the gauge itself can be improved (i.e. to reduce equipment variation EV).

Now that we have a capable measuring system, we now need to think about how we are going to verify that the process is capable and stable. For key parameters (sometimes called special characteristics) an organization should use statistical methods of evaluation.

Ideally we want the variation in the process to be small when compared with the tolerance, and for the process to be centered to the nominal value.

The automotive industry uses process capability indices, known as Ppk or Cpk, to assess the capability of a process. By calculating the process capability, it allows the organization and the customer to predict the amount of parts that the process is capable of producing within the defined specification. For example, for a process capability index of 1.0, the probability is the process would produce 2700 nonconforming parts in every million parts produced (2700ppm), which in the automotive industry would be deemed as totally unacceptable. Many automotive customers are driving suppliers to achieve a process capability index of 2.0, where the probability is the process will achieve a ppm value of less than 1.

Once the first production batch has been produced, and the capability of the process verified, the automotive industry requires suppliers to undertake a product approval process. Many different formats of submissions are specified by customers, but the most common format is Production Part Approval Process (PPAP). The process requires the supplier to collect all the data that demonstrates to the customer that the process is capable of manufacturing parts to their requirements. The supplier submits the information in the form of a warrant.

Information that forms the basis of a submission typically includes:

• Design Record
• Authorized Engineering Change documents
• Customer Engineering Approval
• Design Failure Mode and Effects Analysis (Design FMEA)
• Process Flow Diagram(s)
• Process Failure Mode and Effects Analysis (Process FMEA)
• Control Plan
• Measurement System Analysis Studies
• Dimensional Results
• Records of Material / Performance Test Results
• Initial Process Studies
• Qualified Laboratory Documentation
• Appearance Approval Report (AAR)
• Sample Production Parts
• Master Sample
• Checking Aids
• Customer-Specific Requirements

Once the customer has approved the submission, the organization can then commence ongoing volume production. To ensure ongoing control and capability of the process, the organization needs to ensure that the all the controls defined in the control plan are implemented, which may include monitoring the process by ongoing statistical process control (SPC), and the use of control charts. An example is shown below:

Conclusion

While each of tools outlined above can help ensure that customer requirements can be met, their implementation within a process-based management system should be linked together as shown below:

As product and process changes are proposed, the risks should be evaluated before implementation, and then once approved the appropriate process documentation updated.

Similarly as internal or external complaints or concerns arise, the above documentation should be review and modified as necessary. For example, in the event of a customer complaint, the FMEA should be review to see if the failure mode had been identified, and the appropriate controls were implemented.

In conclusion, for successful implementation that will lead to the goal of continuous improvement, defect prevention and the reduction in variation and waste, people involved not only needs to be trained and be convinced in the benefits of application, but top management need to be committed to providing the resources for applying them.

By telling people to just "fill in the forms", this will never be successful or beneficial.