IFAS – Intelligent Fault Alarm System
IFAS was developed by a Portuguese research group. It is a software implementation of a monitoring and early warning system that can be applied to multiple domains involving multiple sensor data. It has the capability of dealing with imprecision, allowing the system to issue levels of warnings before an error actually occurs. It extracts data from sensors, and semi-automatically transforms and represents it as linguistic variables from nominal behaviours.
Rule-based Systems have been successfully applied in many different fault detection and monitoring technical processes. However, most of these applications are related with decision making (automatic control processes) and not with decision support (human control processes). Further, there is not much work devoted to the application of specialized fuzzy inference concepts and tools to assist human operators in the task of control processes.
We developed an intelligent fault alarm system to monitor faulty behaviors during a drilling process and that includes an inference system for terrain hardness detection, both of which are based on fuzzy logic. Fuzzy logic has the advantage of capturing rules expressed as natural language in a very straightforward manner. It also provides a high degree of flexibility, which allows the system to issue levels of warnings before an error actually occurs (early warning).
The main objective of the fault detection system is to provide a solution for detecting deviations from nominal values in components or processes which are based on automatic sensor data acquisition.
Innovations & Advantages
The novelty of this approach is the combination of a learning module that extracts data from sensors, transforms and represents it as linguistic variables, with a flexible reasoning process that can detect deviations from nominal behaviors and provide alarms ranging from low deviation (green) to critical alarm (red). Moreover, the alarm tool interface is designed as a user-friendly decision support tool.
Another innovation of this approach is the flexible monitoring mechanism, which allows reasoning about deviations, and detect at an early stage any problem encountered in the data being extracted (providing early warning capabilities).
IFAS has been tested in aerospace applications such as ExoMars rover drilling system and gyroscopes degradation in ESA satellites.
Current and Potential Domains of Application
A system like this can be applied whenever there are sensors that extract data, which include some sort of imprecision, that should be measured to detect deviations from nominal values at an early stage, i.e., before any critical problem occurs. This also applies to any equipment that needs monitoring and early fault detection alarms. Potential application areas include energy systems, industrial assembly lines, smart buildings, medical and health, drilling instrumentation, etc.