One of the main problems associated with the transport and manipulation of multiphase flow is the existence of flow regimes, which have a strong influence on important parameters of operation. An example of this occurs in gas-liquid chemical reactors in which maximum coefficients of reaction can be attained by keeping a dispersed-bubbly flow regime to maximize the total interfacial area. Another example is the pneumatic conveying of solids in which the regimes are associated with safety and energy consumption. Thus, the ability to identify flow regimes automatically is very important, specially to maintain multiphase systems operating according to design conditions. This work assesses the use of a self-organizing map (neural network) adapted to the problem of regime identification in horizontal two-phase flows. In order to achieve extensive results, two different types of two-phase flows were considered: gas-solid and gas-liquid. Tests were made to verify the performance of the neural network model, using data collected at the experimental facilities of the Thermal and Fluid Engineering Laboratory of the University of São Paulo at São Carlos. Results show that the neural network is capable of correctly identifying the regimes. The error percentage is bigger when analyzing the same regime with flow rates different from the one used as training data emphasizing the importance of training signals choice.

The main objective of this work is to contribute to the development of a new tomographic reconstruction method well suited for processing signals obtained from electrical or other soft sensing field probes. The adopted approach consists in formulating the reconstruction problem in terms of an error function, which assesses the difference between a prospective and the actual internal contrast distribution (3D image), and searching for its minimum with the help of a specialized genetic algorithm (GA). Numerical simulations have been performed to demonstrate the feasibility of the proposed reconstruction method, as well as to emphasize the relation between the ill-posed nature of the problem and the topology of the minimization hyper-surface, and the importance of considering this relation when designing the numerical solution procedure. Results show that convergence is greatly enhanced when a priori information is introduced in the error function.

The pneumatic conveying of solids in a gas stream is a recurrent process in petrochemical industries. However, due to practical limitations the majority of existing systems have capacities ranging from 1 to 400 tones per hour over distances less than 1000 m, mainly because of a high power consumption per transported unit mass. More specifically, to avoid the formation of dense structures such as dunes and plugs, which, depending on the characteristics of the material and on the availability of a pressure head from the carrier phase may cause a violent pressure surge or a possible line blockage, the system is preferably operated at homogeneous dispersed flow. To sustain such a flow regime high velocities are needed and, accounting for the resulting higher pressure drops, higher power consumption is demanded. An optimized pneumatic conveying system can be conceived with the help of adaptive control techniques. In the context described above, lower transport velocities are allowed if the formation of aggregates that precedes the transition to dense phase flow regimes are automatically detected and destroyed, thus, artificially stabilizing the light phase homogeneous flow regime. This work assesses the reduction in the necessary power that the application of such adaptive control technique could produce. Experimental results are presented for a 45 mm i.d. pneumatic conveying system used to transport Setaria Italica seeds. The instrumentation used to identify the flow regime is constituted of several pressure sensors installed along the transport line. The proposed control strategy is based on processing these signals through a neural network model to assess the flow condition and to mimic an optimized gain scheduled PID algorithm. Preliminary results show that reductions in power consumption can reach 50% when compared with classical non controlled transport.

The formal calibration procedure of a phase fraction meter is based on registering the outputs resulting from imposed phase fractions at known flow regimes. This can be straightforwardly done in laboratory conditions, but is rarely the case in industrial conditions, and particularly for on-site applications. Thus, there is a clear need for less restrictive calibration methods regarding to the prior knowledge of the complete set of inlet conditions. A new procedure is proposed in this work for the on-site construction of the calibration curve from total flown mass values of the homogeneous dispersed phase. The solution is obtained by minimizing a convenient error functional, assembled with data from redundant tests to handle the intrinsic ill-conditioned nature of the problem. Numerical simulations performed for increasing error levels demonstrate that acceptable calibration curves can be reconstructed, even from total mass measured within a precision of up to 2%. Consequently, the method can readily be applied, especially in on-site calibration problems in which classical procedures fail due to the impossibility of having a strict control of all the input/output parameters.

One of the main problems related to the transport and manipulation of multiphase fluids concerns the existence of characteristic flow patterns and its strong influence on important operation parameters. A good example of this occurs in gas-liquid chemical reactors in which maximum efficiencies can be achieved by maintaining a finely dispersed bubbly flow to maximize the total interfacial area. Thus, the ability to automatically detect flow patterns is of crucial importance, especially for the adequate operation of multiphase systems. This work describes the application of a neural model to process the signals delivered by a direct imaging probe to produce a diagnostic of the corresponding flow pattern. The neural model is constituted of six independent neural modules, each of which trained to detect one of the main horizontal flow patterns, and a last winner-take-all layer responsible for resolving when two or more patterns are simultaneously detected. Experimental signals representing different bubbly, intermittent, annular and stratified flow patterns were used to validate the neural model.