Model-predictive control and reinforcement learning in multi-energy system case studies

Ceusters, Glenn; Rodríguez, Román Cantú; García, Ana María Moreno; Franke, Rüdiger; Deconinck, Geert; Helsen, Lieve; Nowé, Ann; Messagie, Maarten; Camargo, Luis Ramirez

doi:10.1016/j.apenergy.2021.117634

Cited by 71 publications

(26 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These trajectories with future forecasts are used in the optimization process to find an optimal control for each time step. However, if such a detailed physical model is not available or the system is too complex to get an accurate model, the optimal control might be difficult to achieve [14]. The MPC relies also on the future forecasts, of which accuracy depends on the used forecasting methods and historical data.…”

Section: Comparison Between Mpc and Rlmentioning

confidence: 99%

“…Modelling and forecasting are two disadvantages of the MPC in the case of the HEMS problem, which have high amount uncertainties and is a complex problem. Additionally, the model-based control does not adapt to the unexpected changes in the real system in the short-or long-term but continue the optimization process based on the applied model [14].…”

Section: Comparison Between Mpc and Rlmentioning

confidence: 99%

“…Additionally, the RL does not require future forecasts because it learns the optimal policy from the direct interaction with the environment [14]. On the other hand, there may exist underlying system dynamics that are not learnt by the RL algorithm despite a large amount of data [14]. This can lead to a lack of optimal control.…”

Section: Comparison Between Mpc and Rlmentioning

confidence: 99%

“…This can be seen as a disadvantage in the cases where historical data is not available for training the controller. When compared to the MPC, as an exploration strategy, the RL rely on a certain amount of randomness, which leads to immature feasibility, robustness and stability theory and constraint handling [14]. These are full-fledged in terms of MPC.…”

Section: Comparison Between Mpc and Rlmentioning

confidence: 99%

See 3 more Smart Citations

A Review – Home Renewable Energy Management Systems in Smart Grids

Kallio

Siroux

2022

IOP Conf. Ser.: Earth Environ. Sci.

View full text Add to dashboard Cite

This paper presents a review on the Home Energy Management Systems (HEMS) for renewable energy production and used optimization methods. The HEMS is an important Smart Grid application. It is used to monitor and optimally manage the energy flows in buildings including renewable energy production, energy storage and smart home appliances. In this paper, two different methods for the optimal HEMS are selected and compared: Model Predictive Control (MPC) and Reinforcement Learning (RL). As a conclusion, the RL method can overcome the disadvantages of the MCP in the highly dynamic environment of buildings and renewable energies, and is a promising method for HEMS in Smart Grids. Finally, an experimental set-up of the hybrid renewable energy system is presented and its operation is discussed under the Time-of-Use energy management strategy.

show abstract

Section: Comparison Between Mpc and Rlmentioning

confidence: 99%

Section: Comparison Between Mpc and Rlmentioning

confidence: 99%

Section: Comparison Between Mpc and Rlmentioning

confidence: 99%

Section: Comparison Between Mpc and Rlmentioning

confidence: 99%

See 2 more Smart Citations

A Review – Home Renewable Energy Management Systems in Smart Grids

Kallio

Siroux

2022

IOP Conf. Ser.: Earth Environ. Sci.

View full text Add to dashboard Cite

show abstract

“…Risbeck et al (2017) extends MPC for a more complex energy system to minimize utility costs. Ceusters et al (2021) developed a mixed-integer linear-MPC to optimize the operation of an energy system with renewable generation (PV), and battery. Other example applications of MPC to building energy systems, we refer to (Avci et al, 2013;Bianchini et al, 2017) for demand-response problems with real-time pricing, (Oldewurtel et al, 2010) for reducing electrical peak demand, (Patteeuw et al, 2016;Qureshi and Jones, 2018) for increasing flexibility and sustainability of energy systems with renewable generation, and (Bianchini et al, 2016;Vrettos et al, 2013) for a pricingformulation.…”

mentioning

confidence: 99%

MPC solution for optimal load shifting for buildings with ON/OFF staged packaged units: Experimental demonstration, and lessons learned

Kim

Braun

2022

Energy and Buildings

View full text Add to dashboard Cite

Energy‐Sharing Economy with Renewable Integration and Management in Communities—a State‐of‐the‐Art Review

Zhou

2024

Adv Energy and Sustain Res

View full text Add to dashboard Cite

Climate change and extreme weather events impose urgent necessities on distributed energy systems with energy flexibility and resilience to survive the district power supply systems. Compared to centralized energy systems, distributed energy systems show more superiorities in power shifting, fast demand response, regional energy allocations, and so on. However, electrification transformation‐ enabled flexible energy sharing, smart energy integration, and advanced management are still in their infancy stages, with unsolved but urgent issues, like large‐scale energy system planning and optimization, energy trading and pricing mechanisms, so on. In this study, energy‐sharing economy with renewable integration and management in communities has been comprehensively reviewed. The “source–grid–load–storage” framework has been implemented on district energy systems with complex relationships among the energy supply–storage–transmission–distribution chain. Roles of energy sharing, integration, and management on energy system sustainability have been provided. Considering widely installed large‐scale renewable energy systems, planning and optimization platforms and tools are provided to guide the distributed/centralized system planning and accurate capacity sizing. Results showed that, multi‐directional power interactions with electrification in buildings and E‐mobility spatiotemporal energy sharing and smart grids enable high renewable penetration, fast power response, energy supply reliability, etc. Unlike traditional fossil fuel‐based power plants, distributed renewable‐supported energy systems (like BIPVs, electric vehicles, etc.) enable emerging energy sharing, integration, and management for energy flexibility and resilience with fast response and energy survivals. Roles of advanced energy include power shifting, fast demand response, regional energy allocations, and so on. Electricity market liberalization can incentivize multi‐stakeholders’ proactivity and market vitality for distributed renewable energy spatiotemporal microgrids. Dynamic energy pricing for the shared power is dependent on the supply–demand relationship, and cost–benefit allocations among different stakeholders need to be investigated for justice and fairness. To avoid performance over or underestimations, advanced approaches are highly necessary for large‐scale energy system planning and optimization. Research results can pave paths for upcoming studies in urban energy system planning, design, and optimization with high energy efficiency and low carbon emissions through joint collaborations from highly economically incentivized participators.

show abstract

Model-predictive control and reinforcement learning in multi-energy system case studies

Cited by 71 publications

References 43 publications

A Review – Home Renewable Energy Management Systems in Smart Grids

A Review – Home Renewable Energy Management Systems in Smart Grids

MPC solution for optimal load shifting for buildings with ON/OFF staged packaged units: Experimental demonstration, and lessons learned

Energy‐Sharing Economy with Renewable Integration and Management in Communities—a State‐of‐the‐Art Review

Contact Info

Product

Resources

About