Self powered neutron detectors are sensors which are used to estimate neutron flux in a nuclear reactor. A typical nuclear reactor has several such detectors. The aim of the project is to develop model and data based techniques for detecting, diagnosing and predicting failures occurring in such detectors to enable optimal maintenance and replacement strategies. Such techniques will involve use of Kalman Filter for model based diagnosis, and use of principal component analysis and fisher discriminant analysis for data based diagnosis. Close cooperation with Bhabha Atomic Research Center (BARC) and Nuclear Power Corporation of India Ltd (NPCIL) is envisaged during the course of the project.