iSixSigma

DFSS: role of QFD, DFMEA and Cause/Effects Matrix

Six Sigma – iSixSigma Forums Old Forums General DFSS: role of QFD, DFMEA and Cause/Effects Matrix

Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
    Posts
  • #66899

    Anna
    Participant

    In implementing Design for Manufacture and Value Engineering techniques, I have made extensive use of QFD and product-function centered Cause/Effect Diagrams to develop categories of Failure Modes and their causes for DFMEA’s (and PFMEA’s) which serve in turn as documentation for the rationale of design changes intended to eliminate or mitigate failure modes. This effort is in complex electro-mechanical products which are usually re-designed incrementally (rarely all-new). The procedure I have used is as follows, but please adjust as needed in your organization to prevent your design teams from having to duplicate work already done.

    Start with the importance-ranked functions the customer (and/or regulators) need the product to do or not do. Transform this to engineering terms (Phase 1 QFD or the first House of Quality). You really only need to do this once per product family unless your customers or the marketplace has changed a great deal, so don’t do it from scratch if your organization has this in it’s records.

    As necessary for your product/industry, transform again to get the actual parameters you will specify in the design and/or control in the process. (Phase 2 or 2 & 3 QFD). Develop a Cause & Effect (Ishikawa) Diagram or Matrix (which one is needed depends on the complexity and interactivity of your system-under-study) for each design parameter which strongly affects the most significant customer needs and wants. Record the results from the C/E analysis on the DFMEA sections listing Failure Modes, Effects, and Causes. Make sure all the failure modes reported by customers and distribution channels have been included. (Review returns, complaints, warranty records, etc.) Finish the FMEA to get an assessment of Severity, Occurance and Detectability.

    Target your product (re)design efforts on places in the DFMEA where the Severity and / or Severity x Occurance are highest. Target your process design & development efforts on items where the Occurance is high or the Detectability Rating is high (detectability is low). Calculate the Sigma value / Z Score of your product using the relative frequency of occurance of the failure modes in the updated-after-redesign DFMEA Feed the results to a team (mostly same people if your organization uses simultaneous engineering) to do the Process FMEA and apply the typical process-oriented Design for 6 Sigma techniques to the Process FMEA results.

    One of the most effective product (re)design techni

    0
    #66900

    Stuart Lamb
    Member

    Thank you Anna for your most comprehensive response. Unfortunately it appears that your reply may have exceeded the maximum size allowed in the message field, or has been truncated for some other reason. I would therefore be much obliged if you could re-post your reply from paragraph five onwards.

    0
    #27366

    Stuart Lamb
    Member

    I am tasked with integrating DFSS into a new product introduction process, the key requirement of which is to understand and reduce risk associated with new products at the earliest stage. Generally, these new products are close variants of existing models for which customer requirements and key issues are well understood. Nevertheless, careless errors creep in causing multiple design iterations, costs etc. We’ve done DFMEA for some time, but (1) with little preparation and (2) ineffective use of the outputs. The result has been that DFMEAs have been long and protracted exercises where many irrelvant possibilities have been considered, and key issues not considered in sufficient detail. I am now looking at using Cause/Effects Matrix as a tool to use before DFMEA to concentrate the DFMEA on the main areas of risk. However, I’m short of examples how this has been used in Product (as opposed to Process) design. I’m also uncertain whether QFD would be better, accepting that this too, whilts clearly similar to C/A matrix, is a protracted process perhaps better suited to new products into new markets. Can anyone offer guidance?

    0
    #67175

    Anna
    Participant

    Stuart –
    Will do, once I get free from the task I have in hand (I’m running a training session for Green Belts at my employers’ site.) We should complete the training by the end of the work day, so I can probably get to this tomorrow.

    I’ll have to re-create it, because I didn’t save it as a file.

    0
Viewing 4 posts - 1 through 4 (of 4 total)

The forum ‘General’ is closed to new topics and replies.