Common Cause Inference is used during fault isolation when diagnostics are generated under the Single-Fault Assumption, when it is assumed that only a single malfunction exists at the time that diagnostics are performed. More specifically, this inference rule assumes that all "failed" tests can be Attributed to a single malfunction. For example, consider the following functional block diagram:
In this sample, tests T1 and T2 have failed and the input to function A is assumed to be good. Application of the Common Cause inference rule removes functions C, D and E from suspicion, since a single failure to one of these functions would not have resulted in both tests failing. (Note: this example assumes that there can be no signal backpropagation when function C or D fails.)
In order for Common Cause Inference to be applied, two failed tests must contain in the their coverage some, but not all, of the functions covered by the other test. In this case, the coverage for T1 consists of function A, B & C, whereas the coverage for T2 is functions A, B, D & E. Notice that functions A & B are in both tests' coverage, whereas function C is covered by only one test (T1) and D & E by the other (T2).
When it is possible that multiple, simultaneous faults exist at the time that diagnostics are performed, application of the Common Cause inference can lead to mis-diagnosis. If, for example, both C and D were to malfunction, then Common Cause isolation would incorrectly conclude that either A or B had failed. To avoid erroneous isolation in multiple-fault scenarios, diagnostics should employ Multiple-Failure, rather than Common Cause, inference.
