“Smart” elements, an implementation point of view

There is a wide range of “smart” element implementations for sensors and actuators. However, there are a number of basic similarities in their implementation. This article presents a common methodology for “smart” sensors and actuators focused on fault diagnosis. The novelty of this work is based upon the integration of different techniques in order to implement fault detection and isolation (FDI) (observers, Kalman filters, discrete Fourier transforms and fuzzy logic). The information processed by the “smart” element is presented as a vector named the symptom vector. This is performed in a similar manner to the SEVA scheme by the use of several values. These values represent: the performance of the element; the evaluation of the device; the evaluation of the measurement. The representation of the information uses a label scale that allows classification of the results from safe to catastrophic conditions. There are three main stages within the processing of “smart” element: capture of information, processing of information and creation of an information standard. In this work, the latter two are studied and integrated using two different implementations, a pressure sensor and a fuel valve. These are evaluated using a number of fault scenarios. A key element in this implementation is the use of fuzzy logic for decision making between different fault scenarios. These are then injected into the dynamics of the evaluated transducer. The use of decision making systems for “smart” sensors and actuators allows the evaluation of information obtained using analytical redundancy