A collaborative training approach for multi energy systems in low‐carbon parks accounting for response characteristics

Abstract

Existing researches lack systematic design of the operation mechanism and overall implementation process of park oriented multi‐state energy system; The existing researches on multi‐state energy systems ignored the dynamic response characteristics of load and the coupling effect of units considering the impact of carbon emissions, and did not go deep into the collaborative management of multiple multi state energy system parks. To this end, this paper first proposes a two‐stage multi‐energy system operation architecture for low carbon parks, and establishes a mathematical model considering carbon emissions, operating costs and other objectives. The response space of load is optimized by considering the dynamic interaction characteristics of load response and the coupling characteristics of unit operation. Then, considering the problems of data security and training efficiency, F‐DDPG technology is used to conduct collaborative training for multiple campus systems. Finally, an example is given to prove the effectiveness of the method, which is not only conducive to reducing the operating cost of the system and the carbon emissions on the supply side, but also improves the convergence of model training. In the case of collaborative training in multiple parks, the method in this paper converges faster and has better economic benefits.

Guest Editorial: Special issue on computational methods and artificial intelligence applications in low‐carbon energy systems

“…for Multi-Energy Systems in Low-Carbon Parks Accounting for Response Characteristics" proposes a horizontal federated reinforcement learning (MEShFRL)-coordinated training method that takes into account the response characteristics of multiple low-carbon parks with a two-stage differentiated incentive mechanism [12]. 4) The paper entitled "Reinforcement Learning-Based Two-Timescale Energy Management for Energy Hub" proposes a deep reinforcement learning algorithm-based twotimescale energy management strategy, where the hourahead policy controls the slow-responding thermal and cooling equipment and the intra-hour policy controls faster-responsive electricity storage equipment [13].…”

“…2)The paper entitled “A Data‐Driven Scheduling Approach for Integrated Electricity‐Hydrogen System Based on Improved DDPG” proposes an improved deep reinforecement learning real‐time scheduling algorithm for an integrated hydropower‐photovoltaic‐hydrogen system to maximize system revenues from the cooperation of various natural resources [11]. 3)The paper entitled “A Collaborative Training Approach for Multi‐Energy Systems in Low‐Carbon Parks Accounting for Response Characteristics” proposes a horizontal federated reinforcement learning (MEShFRL)‐coordinated training method that takes into account the response characteristics of multiple low‐carbon parks with a two‐stage differentiated incentive mechanism [12]. 4)The paper entitled “Reinforcement Learning‐Based Two‐Timescale Energy Management for Energy Hub” proposes a deep reinforcement learning algorithm‐based two‐timescale energy management strategy, where the hour‐ahead policy controls the slow‐responding thermal and cooling equipment and the intra‐hour policy controls faster‐responsive electricity storage equipment [13].…”

Collaborative Optimization Scheduling of Multi-Microgrids Incorporating Hydrogen-Doped Natural Gas and P2G–CCS Coupling under Carbon Trading and Carbon Emission Constraints

“…In Equation (11) The expression for the consumption of carbon dioxide by the P2G unit can be represented as follows [28]:…”

“…P BUY i,max and P SELL i,max are the upper limits of power bought from and sold to the grid in microgrid i. (11) Electrical and heat load constraints The electrical load of microgrid i at time t, denoted by P e i,t is composed of the following: the fixed electrical load, P e,g i,t ; the transferable electrical load, P e,tr i,t ; and the reducible electrical load, P e,cut i,t . The relationship is modeled as follows: .…”

“…Regarding low-carbon policies, Reference [11] established a model focusing on carbon emissions. Based on the carbon trading mechanism, a cooperative model was proposed to determine optimal power trading [12], and Reference [13] established a low-carbon transaction mechanism based on the carbon trading mechanism and the Stackelberg game theory, and Reference [14] developed a trading strategy between multiple entities in the system using the double Stackelberg game framework, focusing on the uncertainty of the carbon trading mechanism.…”