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

TL;DR

This paper compares model-predictive control and reinforcement learning in multi-energy systems, showing RL can outperform traditional methods with proper training, especially when models are imperfect or complex.

Contribution

It introduces a model-free reinforcement learning approach for multi-energy systems and benchmarks it against linear MPC, demonstrating RL's potential to outperform traditional control methods.

Findings

RL outperforms LMPC in simple systems (101.5% vs 98%)

RL achieves higher performance in complex systems (94.6% vs 88.9%)

RL requires extensive training but offers adaptive control without a detailed system model.

Abstract

Model-predictive-control (MPC) offers an optimal control technique to establish and ensure that the total operation cost of multi-energy systems remains at a minimum while fulfilling all system constraints. However, this method presumes an adequate model of the underlying system dynamics, which is prone to modelling errors and is not necessarily adaptive. This has an associated initial and ongoing project-specific engineering cost. In this paper, we present an on- and off-policy multi-objective reinforcement learning (RL) approach, that does not assume a model a priori, benchmarking this against a linear MPC (LMPC - to reflect current practice, though non-linear MPC performs better) - both derived from the general optimal control problem, highlighting their differences and similarities. In a simple multi-energy system (MES) configuration case study, we show that a twin delayed deep deterministic policy gradient (TD3) RL agent offers potential to match and outperform the perfect foresight LMPC benchmark (101.5%). This while the realistic LMPC, i.e. imperfect predictions, only achieves 98%. While in a more complex MES system configuration, the RL agent's performance is generally lower (94.6%), yet still better than the realistic LMPC (88.9%). In both case studies, the RL agents outperformed the realistic LMPC after a training period of 2 years using quarterly interactions with the environment. We conclude that reinforcement learning is a viable optimal control technique for multi-energy systems given adequate constraint handling and pre-training, to avoid unsafe interactions and long training periods, as is proposed in fundamental future work.