Abstract
Fault diagnosis is of great importance especially in the field of aerospace, because spacecrafts are expensive and unique in most cases. Fault diagnosis can not only improve reliability of the spacecrafts but also reduce the workload of ground engineers, time of training astronauts and costs of lunching and running spacecrafts. Constraint inference is an important field in Artificial Intelligence and other areas of computer science. It has been widely used in model-based fault diagnosis, decision support, natural language understanding and other fields. Model-based fault diagnosis is an example of abductive reasoning using a model of the artifact. It has two steps: conflict identification and candidate generation. In the first step, the diagnosis system simulates the system using the model and compares the real observations with the predicted observations of the system. If conflicts are identified, we can get the reasons for the faults in the second step. Therefore, using constraint inference in model-based fault diagnosis, one significant step is building a model to simulate the real system. This article presents a new modeling method and uses this method to realize constraint inference in an example. It has been proved that this method is very convenient and efficient.
