Development of a Fast Gas-Solid Flow Simulation for Control of the Pneumatic Conveying System on Air Seeders
Cousins, Jason D.M. 1993-
MetadataShow full item record
Limitations of the pneumatic conveying system are an obstacle to the improvement of air seeding technology. Operators often run conveying velocities far above the minimum requirement. This is common because lower conveying velocities - which could reduce waste, energy consumption and hydraulic requirements - put the system at risk of blockages and non-uniform distribution. Furthermore, new precision technologies such as variable rate application and sectional control introduce imbalances to the highly coupled and distributed conveying system. Incorporating adaptive control mechanisms has been theorized as a potential means of improving conveying system performance. Real-time prediction of conveying system flow conditions is a prerequisite for the proposed control strategies. There is limited existing research regarding control and modeling for air seeders or similar pneumatic conveying systems. While there is extensive research for multiphase flow modeling, few examples prioritize computational efficiency to the extent that real-time simulation is feasible. Application to control dictates that computational speed, in addition to accuracy, is essential. The purpose of this research was to identify, develop and validate a method for predicting flow conditions within a pneumatic conveying system that is suitable for control applications. A low-computational cost, one-dimensional model and simulation have been developed for fast prediction of bulk multiphase flow conditions within the pneumatic conveying system. The model is a simplified form of the Eulerian-Eulerian (two-fluid) equations for fluid-particle flows. The differential model equations were discretized via the finite volume method and solved using computational fluid dynamics techniques. Specifically, the SIMPLER algorithm for the solution of coupled equations was used. The simulation program, which employs the numerical methods to obtain solutions to the discrete equations, was implemented in MATLAB®. Experimental data were collected using a laboratory apparatus which approximated a straight horizontal pneumatic conveying line. The inner diameter of the experimental conveying line was 57.4 mm. Spherical plastic particles with a mean diameter of 3.56 mm were conveyed. Testing consisted of dilute flows only that were relevant for air seeding conditions. Experiments covered air velocities of 20 to 30 m/s and mass loadings of 0.84 to 4.68. Recorded data included steady-state and transient measurements for fluid pressure and bulk particle velocity. The experimental data were used to validate simulation results. The accuracy of the model for steady-state conditions was acceptable for sufficiently dilute and well-developed flow. The simulation predicted experimental fluid pressure within 6% in all tests. For moderate mass loadings, simulation error for particle velocity was below 10%. At higher mass loadings, accuracy for particle velocity began to deteriorate and an error of > 25% was observed. Analysis of the model’s accuracy for transient conditions was inconclusive. Evidence suggested that transient simulation results may be quite good. However, limitations of the continuous equations and experimental factors complicated the analysis, preventing a definitive verdict regarding transient accuracy. Simulation performance with respect to computing time was excellent. Simulation results were found to be relatively insensitive to the size of time and spatial step used, allowing for the program to execute in less time than was being simulated. The fastest execution recorded required 5.0 sec to simulate 60 sec of transient flow, and results deviated minimally from higher resolution simulations. Results indicated the potential for optimization between speed and accuracy. While the simplified model only calculates a limited number of bulk flow properties, it delivered timely results with reasonable accuracy and with relatively low computational effort. Assessment of the developed model and simulation has concluded a suitable potential for control application. Acceptable accuracy and computing speed were obtained to justify further development efforts. The prescribed methodology provides a foundation for future expansion and improvement. There is potential to incorporate fast multiphase flow simulation into control infrastructure to improve the performance of the air seeder conveying system.
DegreeMaster of Science (M.Sc.)
SupervisorNoble, Scott D
CommitteeBugg, James D; Regier, Christopher N; Evitts, Richard W
Copyright DateJune 2019