An extended self-organizing map with application to the modeling of pulse jet fabric filters

Abstract

Publisher Summary This chapter discusses the modeling methodology based on the framework of local model networks (LMN) to construct an accurate and simplified dynamic model of a pulse jet fabric filters (PJFF). PJFF sometimes called “bag house filters,” have become an attractive option of particulate collection utilities because of the feature that they can meet the stringent particulate emission limits regardless of variation in the operating conditions. Other merits of pulse jet fabric filters are high collection efficiency, on-line cleaning applications, and outside collection that allows the bag maintenance in a clean and safe environment. Despite the recent advances of LMN, a prior knowledge of the processes has to be exploited for determination of the LMN structure and the weighting functions. The chapter describes extended self-organizing map (ESOM) network, which can overcome aforementioned difficulties, developed to construct the LMN using the input–output data. Simulation results of a pulse jet fabric filter example show that the proposed method has better modeling accuracy as compared to both single global model and conventional LMN approach.