In sophisticated and automated technological processes, the consequences of a defect may rapidly spread, resulting in a decrease in process performance or, in the worst-case scenario, a catastrophic collapse. This implies that defects must be identified as soon as feasible, and choices must be taken to prevent their impacts from spreading and to limit process performance deterioration. Various defects influence the behaviour of the process in different ways, and the fault may be identified by ruling out faults for which the anticipated behaviour of the process differs from the actual behaviour. A model represents the anticipated behaviour of the process for various problems in model-based diagnostics.A diagnostic system is a gadget that detects problems. A variety of tests in the diagnostic systems discussed here use observations of the process to verify the consistency of various elements of the model. The collection of tests that is used to determine which problem has occurred must be carefully chosen. Furthermore, using fewer tests reduces the on-line computing cost of operating the diagnostic system and reduces the overall complex and time-consuming task of test creation.A two-step design process for building diagnostic systems is presented, with the ability to choose which tests to employ implicitly by choosing which portions of the model should be tested with each test. The test design for each component may then be done using any current model-based diagnostic method. There are two kinds of design objectives suggested in terms of the capacity to identify defects. The initial aim is to create a sound and comprehensive diagnostic system, which has the following properties. The diagnostic system calculates the flaws that, when combined with the observation, are compatible with the model. The intended isolability is the second objective, and it specifies which defects should be differentiated from other faults.