States Estimation of Pressure Wave Propagating in a Pipe by Particle Filter

Abstract

In these days of increasing demand for condition monitoring of machinery, a detailed estimation of the internal state of a machinery at low cost using a small number of sensors is an important issue. Data assimilation have attracted attention as the methodology to utilize experimental measurements to estimate an internal parameter of numerical model. This study proposes a state estimation method to estimate the state of gas in piping as a state vector in the data assimilation method. The pressure wave propagation in the piping is modeled by concentrated mass model consist of masses, connecting springs, and connecting dampers. The order of the wave propagation model is reduced using modal analysis to improve the accuracy of estimation. A numerical experiment of the one-dimensional sound tube is conducted to confirm the accuracy of estimation. A particle filter algorithm applied to estimate distribution of pressure fluctuation and density of gas in the one-dimensional sound tube model using a few points of measured pressure fluctuation data.