Fault Reconstruction for Dynamic Systems Subjected to Multiplicative Faults

Fault reconstruction can clearly display the size and shape of the fault concerned, which has become one of the important fault diagnosis techniques. Compared with additive faults, multiplicative faults are challenging to reconstruct, which motivates us to develop an effective fault reconstruction technique for multiplicative faults. In this study, for a dynamic system corrupted by unknown input uncertainties and multiplicative faults as well as additive faults, a robust adaptive technique is addressed to reconstruct both multiplicative faults and additive faults, and the system states are reconstructed as well as a by-product. Augmented system approach, unknown input observer and adaptive tracking techniques are united to achieve fault reconstruction. The proposed technique is quite general which is applicable for systems subjected to multiplicative faults only, additive faults only and both multiplicative and additive faults. Moreover, the reconstruction techniques are suitable for both linear and Lipschitz nonlinear dynamic systems. Finally, two engineering-oriented simulated examples demonstrate that additive and multiplicative faults can be reconstructed successfully, proving the effectiveness of the proposed algorithms.