Development of abnormal signal(noise) discrimination method in nuclear power plant

Measuring and characterizing low-level input signals for control systems is extremely difficult in the presence of noise. However, separating the signal from the noise is important for maintaining and controlling the plant safety. This paper describes the experiment made to systematically develop an algorithm(noise discrimination method) which allows for the effective identification and classification of signal and noise. This computerized signal recognition algorithm samples the observed signal at a high frequency in order to recognize the wave pattern and extract characteristics. It has been verified that this is an accurate and highly sensitive algorithm. Thus it is proposed that it would be applicable to radiation monitoring systems for nuclear power plants. (author)