A real-time diagnostic and control technique has been developed for use in electronic circuits whose thermal signature can be correlated to their operating status. Successful implementation of this diagnostic scheme in proof-of-concept experiments required the incorporation of several technological issues into a complete system that has the capability to detect potential fault modes in the system under observation. Included was the ability to: (1) use infrared fiber optics to view components within enclosures and complex geometries, (2) obtain the thermal profile of the system, (3) process and analyze thermal data, (4) implement a simulated artificial neural network to determine the particular condition or fault corresponding to the thermal signature, and (5) perform any necessary corrective action in a timely manner. Infrared optical fibers, routed from individual components to an external array of connectors, were used to collect and transmit energy radiated from those components. An infrared thermal imaging camera was utilized to scan the fiber array and produce an image corresponding to the thermal profile; thus, the thermal signature was obtained in a manner which was neither thermally nor electrically intrusive. Temperature data was then transmitted via an interface bus from the camera system to the control computer where information was converted into a form suitable for input into a trained artificial neural network.; Backpropagation was the neural network algorithm chosen for this application, and it was simulated in software to detect the probable operating condition or fault mode responsible for generating a given thermal image. Inputs to the backpropagation algorithm were the changes in component temperature from a prescribed "normal" operating mode, while the outputs defined the most likely corresponding operating condition. If the system determines that an undesirable fault has occurred, an option could automatically be chosen to modify the operating conditions of the circuit such that component temperatures return to normal and safe values.; The concept was implemented on a capacitor-charging power supply which was adapted to allow implementation of the thermal diagnostic system. For demonstration purposes, twenty component temperatures were monitored via infrared optical fibers, and they were used to detect and distinguish between a total of eleven different fault modes and operating conditions. The diagnostic system successfully detected each of the eleven modes which were generated during the testing phase of the experiments. Requirements for adapting this technique to large-scale systems with hundreds or thousands of components and possibly a hundred different fault modes is also discussed.
展开▼
