Acceso abierto

Research on Reinforcement Learning Based Regulation Scheme for Renewable Energy System in Green Buildings

 y   
21 mar 2025

Cite
Descargar portada

Figure 1.

Photovoltaic power generation in winter
Photovoltaic power generation in winter

Figure 2.

The convergence process of DDPG algorithm
The convergence process of DDPG algorithm

Figure 3.

The operation effectiveness of each system control scheme
The operation effectiveness of each system control scheme

Figure 4.

Simulation results of system regulation
Simulation results of system regulation

Figure 5.

Heat supply and Thermal storage system energy storage level
Heat supply and Thermal storage system energy storage level

Related parameters

Window size/m2 Building heat capacity/(J·K-1) Thermal resistance between indoor and environment/(K·W–1) Ashp heating power/kW
10 7453000 5.28×10–3 2

Operation strategy of heat pump and energy storage

Operation strategy of heat pump
Room temperature Ttin\[T_{\text{t}}^{\text{in}}\]/°C Electrovalence pt(yuan(kW·h)-1)
pt≤0.35 0.35<pt≤0.65 pt>0.65
Low 1 1 0.75
Medium 0.75 0.75 0.5
High 0.25 0 0
Operation strategy of energy storage
Photovoltaic power generation Ttin\[T_{\text{t}}^{\text{in}}\]/kW Electrovalence pt(yuan(kW·h)-1)
pt≤0.35 0.35<pt≤0.65 pt>0.65
Low -1 -1 -1
Medium 1 0 1
High 1 0 1

Robustness of DDPG

Scheme Operating cost(RMB) ATV(°C)
0.9°F 1.8°F 2.4°F 0.9°F 1.8°F 2.4°F
Scheme 1 3023 3034 3029 0.182 0.186 0.235
Scheme 2 2392 2406 2407 0.182 0.186 0.235
Scheme 3 2204.34 2263.91 - 0 0 -
DDPG 2181.3 2284.4 2284 0.003 0.005 0.075