Reconstructive inverse dynamics in feedwater control

摘要

The history of nuclear reactor operations during the last two decades has indicated the need to use advanced control, monitoring, and diagnostics techniques to achieve major improvements in power plant performance. In this paper, the authors present an application of reconstructive inverse dynamics (RID) to the feedwater-train control system for a typical pressurized water reactor (PWR). Existing controllers used in power plants, such as proportional-integral-derivative (PID) controllers, are designed using linear techniques. Although PIDs have proven simple and reliable in reactor operations, their limitation is related to the degree of linearity of the system dynamics. When the system undergoes a state-transition caused by a significant nonlinearity, the linear controller gains may not provide adequate compensation on the system dynamics, regardless of how sophisticated the control designs are. This fact is the main motivation behind searching for an appropriate nonlinear control law. The RID is a version of the inverse-dynamics method currently used in robotics. The power of RID lies in its unique control law for any system with plant nonlinearities. The control law includes inverse dynamics and state reconstruction to follow a demanded trajectory. The RID structure is best understood with a coupled system dynamics.