基于自适应马尔科夫功率预测的混合储能辅助火电机组AGC随机模型预测控制OA

[Objective]With the growing integration of renewable energy sources into the power grid,frequency fluctuations have become a significant challenge.To address this,hybrid energy storage systems(HESS)are used to assist thermal power units in responding to automatic generation control(automatic generation control,AGC)commands.This paper proposes a novel hybrid power distribution strategy based on stochastic model predictive control(SMPC)to enhance the regulation performance of thermal power units in AGC applications,particularly under fluctuating power demands.The aim is to optimize the power allocation between the thermal power unit and the HESS to improve the accuracy,stability,and efficiency of the regulation process,ensuring a more reliable response to AGC signals.[Methods]The proposed strategy first constructs a power demand model for the HESS system,consisting of lithium titanate batteries for high-power storage and lithium iron phosphate batteries for energy storage,based on a Markov probability matrix,which simulates the response of the thermal power unit to AGC commands.An adaptive mechanism is introduced to dynamically adjust the state transition probabilities in real-time,enhancing the accuracy of power demand predictions during AGC fluctuations.Additionally,a scene tree generation method is proposed,which combines probability thresholds with stratified sampling to transform the probability distribution output by the adaptive Markov model into a finite set of scenarios for optimization.This method is designed to better handle the uncertainty of power demand predictions under multiple future scenarios,addressing the inherent variability of AGC command responses.Finally,the strategy integrates the above components into an SMPC controller,which optimizes power distribution between the thermal power unit and HESS in real-time,considering the stochastic nature of power demands and control parameters.[Results]Simulation experiments demonstrate that the proposed strategy significantly outperforms traditional frequency regulation strategies,which do not incorporate power prediction,and static SMPC strategies that lack dynamic correction of state transition probabilities.The performance index Kp is improved by 14.1%and 7.5%,respectively,showing that the SMPC strategy with adaptive power demand forecasting can achieve more precise and stable regulation performance.Additionally,the model's ability to handle uncertainty in power demand prediction allows for more accurate and timely responses to AGC fluctuations,resulting in better coordination between the thermal power unit and the HESS.[Conclusion]The proposed strategy effectively enhances the collaborative regulation performance between the thermal power unit and HESS,offering strong application potential.Further optimization of the model can improve its robustness and adaptability in practical applications,advancing the implementation of this technology.