At present

the modeling research on combined heating and power(CHP) system of ‘gas internal combustion engine + lithium bromide absorption type hot and cold water unit' mainly focuses on mechanism modeling. However

using a data-driven approach to model the system is easier and more accurate. In this paper

by analyzing the system structure

process flow and dynamic characteristics

the input initial auxiliary variables of the system are selected

and the main input variables are determined by principal component analysis. The system is simplified into a three-input and three-output multi-variable coupling system. Then

by using particle swarm optimization algorithm based on dynamic inertia weight and acceleration factor(PSO-DW-DAC)

the operating data under heating conditions are identified and modeled. The verification shows that

the PSO-DW-DAC has better search ability and higher search efficiency than the conventional particle swarm optimization algorithm(PSO)

and it is feasible to model the cogeneration system through data-driven methods. The result provides a basis for system control and optimization operation research.