The objective of this work is to experimentally examine the feasibility of using artificial neural networks (ANNs) for the identification of the flow regimes in three phase gas/liquid/pulp fiber systems, using the statistical characteristics of pressure fluctuations measured by a single pressure sensor. Experiments were carried out in a transparent, vertical circular column with 1.8 m height and 5.08 cm inner diameter, using mixtures of air, water, and Kraft softwood paper pulp. The pulp consistency (weight fraction of pulp in the pulp/water mixture) was varied in the low consistency range of similar to0.0-1.5%. Flow structures and other hydrodynamic characteristics were experimentally investigated, using visual observation, gamma-ray densitometry, and flash X-ray photography. Local pressure fluctuations at a station 1.2 m above the test section inlet were recorded using a high-sensitivity pressure transducer that was installed in a manner not to cause flow disturbance. Various statistical characteristics of the pressure fluctuations, as well as the characteristics of their autocorrelations, were examined as potential input parameters for ANNs trained to recognize the flow regimes solely on the basis of these statistical pressure fluctuation characteristics. Three-layer, feed-forward ANNs could learn to identify four major flow patterns (bubbly, plug, churn, and slug) well, using the standard deviation, coefficients of skewness and kurtosis, and several time shift autocorrelations of normalized pressure signals as input. The results indicate the feasibility of using ANNs for flow regime identification in three-phase gas/liquid/pulp flow systems.