A Control Strategy of Thermal Power Unit Based on AGC Prediction Command

The uncertainty and volatility of new energy power generation itself is bound to affect the security of power grid. In the case of enough peak load and frequency regulation capacity of thermal power units, by means of changing the output of thermal power unit the fluctuation in power grid can be stabilized, and the impact caused by grid-connection of new energy power units can also be mitigated. To ensure secure, economic and stable operation of power grid and to make thermal power units possessing enough regulation capability, an AGC command prediction-based feedforward predictive control method for thermal power units was proposed. Firstly, a BP neural network model was constructed to predict the output of wind farm and photovoltaic (abbr. PV) power station as well as the load of power grid to obtain the source load analysis-based plant level AGC prediction curve. According to this prediction command the load of thermal power plant was distributed, and the obtained variable load trend was taken as the feedforward information led into the control system of thermal power units to revise the AGC command curve of thermal power units, and to perform the optimal feedforward predictive control of the units. It is verified that the proposed control method can make thermal power units responded in advance, so the response capability of the units can be effectively improved. Comparing with conventional control methods, the proposed method improves both control accuracy and response efficiency to a certain extent, thus the regulation capacity of thermal power units is enhanced.