Loading...

Model Predictive Control Applications in Turbocompressor Control System

Sheikhbahaei, Reza | 2019

199 Viewed
  1. Type of Document: Ph.D. Dissertation
  2. Language: Farsi
  3. Document No: 55095 (08)
  4. University: Sharif University of Technology
  5. Department: Mechanical Engineering
  6. Advisor(s): Alasty, Aria; Vossoughi, Gholamreza
  7. Abstract:
  8. The purpose of this study is to develop model predictive control approach to control a turbocompressor. The turbocompressor which is consisted of a gas turbine and a centrifugal compressor is manufactured for providing the necessary pressure for gas transport via the gas transmission network. The turbocompressor control system regulates the compressor outlet pressure by controlling the rotational speed of the gas turbine, while preventing the surge phenomenon in the compressor. Surge is a serious instability in the compressor that is associated with periodic fluctuations in system variables and can cause heavy damage to expensive machine parts.In this research, model predictive control has been used both in controlling the rotational speed of the gas turbine and in predicting surge and its prevention. The problem of surge prevention has been investigated under the assumption that the anti-surge valve is faulty, and for this purpose a fault-tolerant explicit model predictive controller has been designed.To design this controller, explicit model predictive control theory has been extended so that a new algorithm is proposed to control a constrained system under the influence of disturbances, uncertainties and actuator faults. This algorithm is based on dynamic programming, constraint rearrangement and multi-parametric programming. First of all, the dynamic programming is used to recast the problem as a multi-stage optimization problem. Afterwards, the constraints are rearranged in an innovative manner to take into account the worst admissible situation of unknown bounded disturbances, uncertainties, and actuator faults. Then, the explicit solution of the reformulated optimization problem for each stage is obtained using the multi-parametric programming approaches. Finally, a recursive procedure for combination and substitution of the solutions is presented to extract the desired explicit control law.In order to control the rotational speed of the gas turbine, a thermodynamic model for its nonlinear behavior was first developed. This model is then tuned during an optimization process using an extended Kalman filter. After that a hybrid model is extracted by which the explicit model predictive controller has been designed. To prevent surge in a compressor with a faulty anti-surge valve, an analytical model of the compressor is developed first and the explicit model predictive controller is then designed on the basis of the proposed control theory. The simulation results in different sections of the dissertation show the acceptable performance of the proposed control system.
  9. Keywords:
  10. Centrifugal Compressor ; Dynamic Modeling ; Robustification ; Surge (Fluid) ; Fault-tolerant Control ; Explicit Model Based Predictive Controller (MPC) ; Gas Turbines ; Twin Shaft Gas Turbine

 Digital Object List

 Bookmark

No TOC