Abstract: Dynamic thermal performance of a district heating system can be utilized to improve the peak regulation capacity of combined heat and power units so as to enhance wind power integration in heating seasons. Modeling dynamic thermal performance with the physical model involves individual characteristics of each part inside the system and topological structure characteristics of pipelines network， resulting in that the combined heat and power dispatch model based on the physical model is complicated， which is difficult to be applied in large-scale power grid analysis. Focusing on the system overall performance， a reduced-form regression model is built to describe the dynamic thermal performance of the district heating system under the quality regulation mode by utilizing the auto-regressive distributed lag time series. The model connotation is analyzed and the modeling method is given. The combined heat and power dispatch model based on the regression model is built. Compared with the dispatch mode based on physical model， the dispatch mode based on the regression model can reduce variable number and calculation time， which has good simplification effect.

Key words: district heating system, dynamic thermal performance, regression model, combined heat and power dispatch, wind power integration