|
Part 1 of this series dealt with overall wiring system design as providing the critical foundation for the growth in automotive electronics systems and functionality. In that installment we looked at how to develop a set of system diagrams into a specific vehicle architecture, taking into account the complexity of a modern vehicle (both in terms of components and the option/variant complexity arriving from increased customer choice). We are now going to look at how the designs, at various stages of maturity, can be analysed and verified.
Simulation and analysis techniques can be applied to an electrical distribution design flow to make it more efficient and less costly. Critical elements of such methods include early system simulation, advanced interpretive analysis (such as sneak-circuit testing), Failure Modes & Effects Analysis (FMEA), and rules driven methodologies for component sizing.
The big picture
Despite the cost of vehicle recalls, or the cost of grounding aircraft and rolling stock, analysis is rarely given a high priority. This means that such methods are often left until late in the design cycle, which perversely means that there are greatly reduced benefits of the analysis.
This is perhaps difficult to understandwhy do it so late when costs of change are so high (see figure above (top)), when the benefits of doing it early are so great (figure above (bottom))? The answer is surprisingly simplemostly it's not analysis being performed, it's simulation, and there is a real difference
View a full-size image
The figure above highlights the difference between simulation and analysis. Simulation just produces a set of (often numerical) results. Those results must then be interpreted and conclusions drawn.
Consider our two engineers in the diagram; each is seeing the same results, but they have very different perspectives. Even worse the cost sensitive engineer on the right has made a change (sharing a ground) that has introduced a failure effect that disables every airbag in the vehicle when a single splice fails.
Most organizations do not have time to place a large team on analysis, and even if they are able to, cannot ensure that a consistent set of rules are applied for every design.
What is really needed are analysis tools which automatically extract the results of simulations and interpret them for the user, enabling them to make the right decisions at the right time.
The remainder of this article illustrates how, and when, such tools can be deployed.
|