
Case Study: Application of Algorithm DFMEAs
Failure Mode and Effects Analysis (FMEA) is a tool to analyze potential risk in
product designs, manufacturing processes, and custom built machinery.
PlanTech developed a methodology to use the FMEA tool to analyze software / algorithms
by comparing outputs of an algorithm against the inputs.
An algorithm is a logical entity designed to perform a specific function or set
of functions. It may consist of one or more subroutines. It receives logical inputs,
which it uses to calculate the desired output(s).
Inputs can be via sensors or multiple other algorithms commonly called rings.
Types of Algorithms:
Control algorithm
An algorithm the primary purpose of which is to report data from one or more sensors,
or to control some electromechanical function of the vehicle.
Diagnostic algorithm:
An algorithm the primary purpose of which is to detect and report an error in another
data item and / or the sensors which determine the data item. For example, in automotive
diagnostic algorithms, before the catalyst monitoring diagnostic can return reliable
results, the engine must be warmed up, so coolant temperature is checked. The inputs
used for these checks are termed enable criteria. This distinction is especially
useful in dealing with algorithms that have a lot of enable criteria and more than
one output
PlanTech worked with the Powertrain division of an OEM facilitating Algorithm Failure
Modes and Effects Analysis (AFMEA) on a software ring pertaining to Performance
Launch Control (PLC).
The subject matter experts who attended were a Global Technical Specialist, two
Algorithm Engineers, a Service Engineer, and a Software Calibrator. The AFMEA was
initiated well in advance of the design release date. During the subsequent meetings
the potential failure modes discovered and mitigated with action items would have
resulted in large warranty costs if not caught ahead of time within the AFMEA. In
addition, the AFMEA meeting was the only time this array of engineers was assembled
for discussion regarding the design of the PLC ring. Normally they work individually
with minimal interaction. This communication enabled the critical discussion to
occur that streamlined the design and headed off downstream warranty costs. The
entire group commented at least once per meeting how valuable the AFMEA exercise
was as a part of their design development process.
The results:
- Potentially large warranty costs were avoided.
- The design was created more efficiently with less code needed for the algorithm.
- Valuable engineering time was saved by not having to re-work a bad design.
PlanTech is a division of Omnex Inc.
Omnex is an international consulting, training and software development organization
specializing in management system solutions in Quality, Environmental, and Health
and Safety standards-based management systems. Omnex also leads the way with Lean,
Six Sigma and other breakthrough systems and methods of performance enhancement,
supported by Omnex Systems, LLC, software solutions for Enterprise Wide Quality
Management Systems®.

|