A colleague recently wrote to me regarding the merits of using statistical energy analysis to predict the vibration levels in a NASA launch vehicle.
Here is my reply:
I use “old-school” empirical scaling techniques. These work reasonably well for deriving component-level maximum predicted environments.
I consider statistical energy analysis (SEA) to be more of an academic tool. It requires many assumptions regarding external acoustic pressure field type, coupling loss factors, modal density, impedance, radiation efficiency, critical and coincident frequencies, distinguishing between acoustically fast and slow modes, etc.
Commercial SEA software users tend to gloss over all of these parameters and just use whatever default values are buried in the code.
So I would prefer to use SEA as a secondary tool.
But please keep mind one all-encompassing truth….
The process of deriving environments by whatever means is the process of building a justification story for one’s own engineering judgment of what those levels should be, heavily weighted with past experience with empirical data.
There is no sarcasm in the previous statement.
– Tom Irvine