Reliability and risk analysis using artificial neural networks
- Sandia National Labs., Albuquerque, NM (United States)
This paper discusses preliminary research at Sandia National Laboratories into the application of artificial neural networks for reliability and risk analysis. The goal of this effort is to develop a reliability based methodology that captures the complex relationship between uncertainty in material properties and manufacturing processes and the resulting uncertainty in life prediction estimates. The inputs to the neural network model are probability density functions describing system characteristics and the output is a statistical description of system performance. The most recent application of this methodology involves the comparison of various low-residue, lead-free soldering processes with the desire to minimize the associated waste streams with no reduction in product reliability. Model inputs include statistical descriptions of various material properties such as the coefficients of thermal expansion of solder and substrate. Consideration is also given to stochastic variation in the operational environment to which the electronic components might be exposed. Model output includes a probabilistic characterization of the fatigue life of the surface mounted component.
- Research Organization:
- Pacific Northwest Lab., Richland, WA (United States)
- DOE Contract Number:
- AC04-94AL85000
- OSTI ID:
- 377052
- Report Number(s):
- PNL-SA--26375; CONF-9503142--; ON: DE96009360
- Country of Publication:
- United States
- Language:
- English
Similar Records
System fatigue reliability and inspection planning for offshore platforms
Reliability fatigue design of welded structures