Download files
Abstract
This thesis aims to develop a practical advanced control solution for coal-fired power plant.
Although the power plant is a highly coupled nonlinear thermodynamic multivariable process, it is still controlled by conventional control strategies. The current SISO control strategy
is designed for base-load operation, and has exhibited poor performance against load change transients. Though many studies have suggested various advanced controls, few of them have been taken on by the power industry.
The author looks at
the physical and engineering restrictions that may have prevented advanced control implementation in power plant, and combines theoretical, analytical and practical knowledge to workout a practical control strategy, which is an advanced control solution for improving plant performance.
A UNSW nonlinear simulator is developed by the author, which represents the dynamics of the power plant process very well, including its conventional controllers. Through system parameter settings and controller turnings, the simulator is fitted to plant data as a benchmark power plant model for control study.
Multivariable control is a natural choice for the highly coupled power plant system. The challenge is to work out a control design method and procedure, where the resulting controller not only significantly improves the performance, but at the same time meets the constraints of current plant control system and operation conditions. The Non-Minimal State Space (NMSS) method has the advantage of using only measured input and output signals of the process, and the Linear Quadratic Regulator (LQR) has proved effective for highly coupled processes, while the multivariable Proportional-Integral-Plus Forward Pass(PIP-FP) control structure is well suited for this control application. Based on these well founded control theories and methods, the author presents a practical control design and a design procedure to implement the advanced controller.
Extensive simulation on the nonlinear power plant simulator shows that the NMSS/PIP-FP multivariable control can significantly improve the performance under nonlinear wide load change conditions. The author hopes that the present methods and the associated design procedure will enable power plant engineers to carry out the control development systematically, and implement the advanced control on a real plant in the near future.