Monitors
To reduce exposure times, dedicated hardware circuitry, software code, or various test methods are often used to detect latent failures. These additional failure detection methods are referred to as monitors. Because a monitor might not detect 100 percent of failures, you indicate the percentage of failures that it detects. This percentage is known as the failure detection percentage or failure detection coverage.
In the FTA module, the default value for Failure Detection Percentage is a number sign. If you do not enter a value in this field, a 0 value is used to indicate that no monitor is being used to detect a failure prior to a check interval. If you enter a value between 0 and 100, there are two cases when the item is failed:
• The failure is detectable by the monitor for the percentage specified for Failure Detection Percentage.
• The failure is not detectable with the remaining time. In other words, the item failure is not detectable by the monitor with a percentage value equal to (100 - Failure Detection Percentage).
If an item failure is detectable, then the failure is detected by the monitor given that the monitor is working properly. However, if the monitor is subject to failures, the monitor might not be available when it is needed. To model a perfect monitor, which is available at all times, you select Self Checking for Detection Method. Mathematically, this selection is equivalent to the case where the detection method is monitored and the monitor failure rate is equal to zero. Further, if you also enter 100 for Failure Detection Percentage, then this is mathematically equivalent to saying that the failure of the item is not latent.
When the monitor is subject to failures, it is a common practice to verify the status of the monitor. In the FTA module, you indicate the periodic interval at which the monitor operation is verified for Scrub Time. Because the monitor scrub verification itself might not be perfect, you can indicate the percentage of time that the verification detects and corrects monitor failures for Scrub Verification Percent. Finally, you specify the time at which a latent failure of the monitor is corrected for Monitor Check Interval.
When the monitor is failed at the time of the check, then the item that is monitored might also be in a failed stated. Therefore, the item is also verified if the monitor is in a failed state. Hence, the check interval for the item that is being watched by the monitor is less than its own check interval. Similarly, when the item is found failed at the time of its maintenance check, it indicates that the monitor is already failed. Hence, the monitor is also corrected. Therefore, at the time of the item's maintenance check, both the item and the monitor are verified.
 |
When a fault tree is build using both high-level and explicit modeling features, only worst case probability calculations are applicable in most cases.
|