【目的】 新能源的随机波动性对新型电力系统的安全稳定运行造成了挑战，电蓄热资源作为消纳波动性新能源的优质资源，已在我国“三北”地区大规模应用。蓄热介质参数值受其温度影响，电蓄热装备在实际运行中会发生蓄热参数漂移现象，导致对其调控偏离预期。鉴于此，提出一种计及蓄热体参数变化的电蓄热装备控制策略。【方法】 首先建立了一种考虑蓄热传递过程的电蓄热装备参数化动态模型，然后针对蓄热参数漂移问题提出了一种基于参数投影的模型参数在线辨识算法；在此基础上，构建了测-辨-控架构下计及蓄热体参数变化的新能源电蓄热协同调控模式，设计了基于动态模型在线自更新的电蓄热装备自适应模型预测控制算法，自动匹配电蓄热装备模型参数的时变特征。【结果】 算例分析结果表明，所提方法能有效应对电蓄热装备的参数漂移，与传统模型预测控制相比，所提自适应模型预测控制将蓄热体温度预测均方根误差由23.22 ℃降低至1.06 ℃，降幅达95.4%；系统日购电成本由431.32元降低至341.45元，降幅为20.8%。【结论】 该方法在控制精度、经济性等方面较传统方法具有明显优势，能够为高比例新能源环境下电蓄热的灵活调控和经济运行提供有效支撑。

[Objective] The stochastic volatility of renewable energy sources presents challenges to the safe and stable operation of the new-type power system. Electric thermal storage systems have emerged as high-quality resources for accommodating variable renewable energy generation and have been widely deployed across the Three-North regions of China. However， the parameters of heat storage media vary with temperature， causing parameter drift in electric thermal storage equipment during operation and resulting in deviations between actual regulation performance and expected outcomes. Accordingly， this paper proposes a control strategy for electric thermal storage equipment that explicitly accounts for parameter variations in the heat storage medium. [Methods] First， a parametric dynamic model of electric thermal storage equipment is established considering the heat transfer process. Subsequently， an online parameter identification algorithm based on parameter projection is developed to address the parameter drift. Building upon this foundation， a coordinated regulation framework for renewable energy and electric thermal storage is constructed within a measurement-identification-control architecture that considers parameter variations of the heat storage medium. An adaptive model predictive control algorithm with online self-updating of the dynamic model is designed to accommodate the time-varying characteristics of model parameters of electric thermal storage equipment. [Results] Numerical examples verify that the proposed method effectively mitigates parameter drift in electric thermal storage equipment. Compared with the traditional model predictive control， the proposed adaptive model predictive control reduces the root mean square error of heat storage temperature prediction from 23.22 °C to 1.06 °C， a reduction of 95.4%. The daily power procurement cost of the system drops from CNY 431.32 to CNY 341.45， representing a decrease of 20.8%. [Conclusions] The proposed method demonstrates superior performance in control accuracy and economic efficiency compared with conventional approaches. It provides effective support for flexible regulation and cost-effective operation of electric thermal storage systems under high renewable energy penetration.