Some thoughts on the advanced control of electric power plants

Abstract

Over the last decade, there has been considerable change in the electric power generation industry. Larger, faster plants have been constructed and old plants have been forced to operate in a much more dynamic environment, thanks to strong competition and market forces. Massive changes have also occurred in information technology, designed to improve power plant performance, by including large amounts of computing and digital circuits, together with very advanced man-machine interface tools, allowing efficient information retrieval and display at reasonable cost. By contrast, the control systems implemented in the distributed control computer systems remain largely conventional control, a mix of various types of PID control and feedforward, which is not at all what is thought of as advanced control. This paper looks at this issue, and describes some of the advanced controllers that might be very effective in the industry, including many that are being offered to the industry by the control vendor companies. The paper then tries to give reasons why these systems have not been taken up in any significant way, and by using examples from our own experience, attempts to suggest how the situation might be improved.

Performance Design of Image-Oxygen Based Cascade Control Loops for Boiler Combustion Processes

“…2,3 For an overview of applications of advanced control in power generation and industrial energy, see the survey paper. 4 However, the biggest problem in the conventional control strategies mentioned is that control systems based on the oxygen content come in with a certain delay that causes final product loss. When the disturbances, such as a change in the inlet air temperature or in the inlet fuel flow rate, enter the combustor, any of these disturbances will affect the outlet oxygen content of the flue gas.…”

The Use of the Åström-Bell Model for the Design of Drum Level Controllers in Power Plant Boilers

“…In the boiler area, many models now exist ranging from complex knowledge based models to experimental models derived from special plant tests. In the middle of this range are so-called 'interpretation' models [Maffezzoni (1997), Rees and Lu (2002)]. These models are complex enough to capture the essential physics whilst at the same time have good control design features.…”