摘 要：为提升高比例新能源大量并网时热电联产机组的调峰经济性与运行灵活性，提出了一种源荷协同优化方法，通过融合建筑热惯性、管网延迟与动态煤耗计算方法，结合混合自适应粒子群优化算法，构建了一体化优化模型，以实现热负荷精准预测与经济分配。以某电厂2×350 MW机组为实证对象，基于EBSILON建立机组数字孪生模型，提取变工况数据并构建计及管网热动态特性的负荷预测模型以及煤耗动态特性模型，采用融合非线性时变惯性权重与拉丁超立方初始化的改进粒子群算法，以系统净收益最大为目标实现源荷协同优化。仿真结果表明：所提方法在不仅能够提升机组变工况运行经济性，还能够在增强系统调节稳定性的同时，标准煤耗率较传统分配方法降低3.5g/(kW?h)，单日净收益增加17.7万元。同时，依托机组煤耗动态模型与管网动态特性的协同作用，系统通过匹配机组最优运行曲线进一步实现日煤耗降低约4.19t，充分验证了所提方法对提升电厂综合效益的有效性

Abstract: To improve peak-shaving economics and operational flexibility of CHP units under high-penetration renewable energy integration, this study proposes a source-load collaborative optimization approach. An integrated framework is developed by incorporating building thermal inertia, pipe network delay, dynamic coal consumption calculation, and a hybrid adaptive PSO algorithm, for precise heat load forecasting and cost-effective allocation.Taking two 350 MW CHP units as the case study, a digital twin model is built via EBSILON. Off-design data are used to construct a load forecasting model (accounting for heating network thermal dynamics) and a dynamic coal consumption model. An improved PSO (with nonlinear time-varying inertia weight and Latin hypercube initialization) is applied to optimize source-load coordination for maximum net profit. Simulation results show that the proposed method can not only improve the economic efficiency of unit off-design operation, but also enhance the system regulation stability—while reducing the standard coal consumption rate by 3.5 g/(kW·h) and increasing the daily net income by 177,000 yuan compared with the traditional allocation method. Further, synergies between the dynamic coal consumption and heating network models enable an additional 4.19 t daily coal reduction by matching optimal unit operation curves, validating the method’s effectiveness in improving power plant comprehensive benefits.