Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.605983
Title: Distributed nonlinear state-dependent model predictive control and estimation for power generation plants
Author: Abokhatwa, Salah G.
ISNI:       0000 0004 5359 8798
Awarding Body: University of Strathclyde
Current Institution: University of Strathclyde
Date of Award: 2014
Availability of Full Text:
Access through EThOS:
Access through Institution:
Abstract:
Centralized model predictive control (MPC) is often considered impractical, inflexible and unsuitable for controlling large-scale systems due to several factors such as large computational effort and difficulty to meet all operational objectives. Therefore, industrial large-scale systems are usually controlled by a distributed control framework. In this thesis, novel sequential nonlinear Distributed Model Predictive Control (DMPC) algorithms for large-scale systems that can handle constraints are proposed. The proposed algorithms are based on nonlinear MPC strategy, which uses a state-dependent nonlinear model to reduce the complexity of solving optimization problem. In this distributed framework, the overall system is divided into several interconnected subsystems and each subsystem is controlled by local MPC. These local MPCs solve convex optimization problem and exchange information via one directional communication channel at each sampling time to achieve the global performance. The proposed algorithms are applied to an industrial power plant model to improve power generation efficiency. A non-linear dynamic model of Combined Cycle Power Plant (CCPP) using the laws of physics was first developed and simulated using decentralized PID controllers. Then, a supervisory controller using linear constrained MPC was designed to tune the performance of the PID controllers. Next, a supervisory centralized nonlinear model predictive control (NMPC) algorithm based on state-dependent models was developed to control the nonlinear plant over a wide operating range. Finally, two sequential DMPC algorithms based on state-dependent models were developed. The lack of states measurement were handled by designing nonlinear distributed state estimation algorithms using state-dependent differential Riccati equation (SDDRE) Kalman filter. Numerical simulation results show that the performance of the proposed DMPC algorithms is close to the centralized NMPC but computationally more efficient compared to the centralized one.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.605983  DOI: Not available
Share: