Fuzzy modeling and fast model predictive control of gas turbine system. (1st June 2020)

• Record Type: Journal Article
• Authors: Hou, Guolian; Gong, Linjuan; Huang, Congzhi; Zhang, Jianhua
• Is Part Of: Energy. Volume 200(2020)
• Publication Date: 2020-06-01
• Subjects: Gas turbine system -- Fuzzy model -- Entropy-based clustering -- Subspace identification -- Fast model predictive control; Power resources -- Periodicals; Power (Mechanics) -- Periodicals; Energy consumption -- Periodicals; 333.7905
• DOI: 10.1016/j.energy.2020.117465
• Languages: English
• ISSNs: 0360-5442

Abstract

Abstract: In terms of the inner complex characteristics and fast dynamic of the gas turbine system, it is a great challenge to realize the safe, stable, and efficient operation of gas turbine system in power generation process like combined cycle unit (CCU). For the purpose of achieving high tracking performance and disturbance rejection ability within less settling time under various operating conditions, an improved fuzzy modeling approach and corresponding fast model predictive control (Fast-MPC) algorithm are introduced and applied to a gas turbine system. Considering the importance of an accurate system model in the performance assurance and promotion of controller, a fuzzy modeling technique adopting the entropy-based clustering and subspace identification (SID) is presented to identify the model of gas turbine system at first. The clustering process can realize the automatic determinations of cluster number along with cluster centers. Furthermore, on account of the incremental data around each cluster centers, the SID method is utilized for the acquisition of state-space model required in control process. Then, in the Fast-MPC, the optimization variables of original quadratic programming problem are reordered reasonably, which different from the conventional MPC. Afterward, an improved original obstacle internal point method combined with warm start strategy is employed for the achievement of higher computational efficiency. Finally, extensive simulation experiments are implemented to verify the remarkable accuracy of the identified model, and the excellent settling rapidity of the designed control algorithm. Highlights: A fast model predictive control algorithm is proposed for high control performance. A modeling method combined entropy clustering with subspace identification is adopted. The proposed modeling and control methods are easy to put into practice. The high modeling precision and excellent regulating speed of the Fast-MPC are verified.