Real-time Energy Management for the Integrated Heat and Power System Using Approximate Dynamic Programming

This paper proposes the real-time energy management for the integrated heat and power system (IHPS) based on the approximate dynamic programming (ADP) algorithm. Considering the energy storage and the heat loss in the pipelines, the multi-time period optimization is formulated. The uncertainties brought by the wind power, electricity price, electric load, and heat load, are also taken into account. The proposed ADP algorithm can obtain the near-optimal solution by recursively solving the Bellman’s equation, meanwhile update the real-time forecast information to improve the optimality under the stochastic environment. Simulation results demonstrate the superiority of the ADP algorithm compared to existing real-time decision-making methods such as model predictive control (MPC) in both deterministic and stochastic cases.