Classical and parametric spectral analysis instrument using a visual programming language

Tube leak detection in fossil-fueled power plant boilers has traditionally used classical power spectral density estimates based on fast Fourier transform methods. This approach has achieved mixed results due to the complex signal environment present in the boiler. The authors compare the results of classical techniques with another approach based on parametric power spectral estimation methods. It is shown that autoregressive-moving-average methods sometimes result in earlier detection of tube leaks. With the advent of visual programming languages for digital signal processing systems, the development time required to implement a parametric spectral estimation system is greatly reduced compared to that for traditional text-based programming environments.<
>