Introduction. In the paper a fault isolation problem in the devices combining digital unit by functional diagnostics methods is considered. Networks of state automates are accepted as mathematical models of the devices. Assumed, that functional diagnostics devices for each network component are preliminarily constructed in an optimal way and they consist of a control automata and of a fault discriminator of unit dimension.Aim. To develop functional diagnostics method based on theoretical analysis allowing to decide fault isolation problem in networks of state automation and to reduce computational complexity and hardware redundancy.Materials and methods. An analysis of mathematical description of a network of state automation and functional diagnostics devices for each network component was presented in terms of algebraic theory of functional diagnosis of dynamic systems. A possibility to transform the set of known functional diagnostics devices of the network was demonstrated. The possibility provided a localization of the network component with an error, if the component was unique.Results. A searching procedure of the analytical equations determining supervision automata and fault discriminator for the whole network was proposed. The case when initial functional diagnostics devices for each network component were defined by scalar functions was considered. The obtained result was generalized to the case, when mentioned devices were defined by vector functions. The application of the described method was demonstrated in the example of construction functional diagnostics devices for simplified fragment of the device for forming priorities of mutual aircraft navigation system.Conclusion. Estimation of results by an order criterion was obtained. It was established that with an increase in the number of network components, the reduction of intentioned redundancy by functional diagnostics devices compared with the original version increased significantly.
Fault tolerance for industrial actuators in absence of accurate models and hardware redundancy
