Circuit Reliability Assessment Using Failure Distribution Models

The patent changes the assessment parameters from binary failure/non-failure to a continuous distribution of failure effects. By modeling component failures with probability distributions and simulating their varying impacts on circuit parameters (such as delay, power consumption, and signal integrity), the assessment transitions from conservative worst-case estimates to more accurate statistical evaluations that reflect real-world variability.

The patent applies partial action by considering only those component failures that actually impact circuit performance below a threshold, rather than assuming all failures are catastrophic. By simulating partial failure effects and filtering out non-critical variations, the method avoids over-conservatism while maintaining reliability assessment accuracy.

The patent performs preliminary actions by pre-characterizing component failures through testing and modeling before circuit assessment. Failure modes are identified, parameter distributions are established, and impact models are developed in advance, allowing the actual reliability assessment to use these pre-computed models rather than performing complex real-time simulations.

The patent uses copying by creating simplified models that replicate the essential behavior of failed components. Instead of simulating every possible failure mode in detail, the method creates representative models of failure effects (such as parameter shifts or performance degradation patterns) that capture the essential impact on circuit operation, reducing complexity while maintaining accuracy.