Stochastic Optimization of Braking Energy Storage and Ventilation in a Subway Station

Tristan Rigaut, Pierre Carpentier, Jean‐Philippe Chancelier, Michel De Lara, Julien WaeytensIEEE Transactions on Power Systems 2018 · DOI

TL;DR

This paper introduces a smart energy management system for subway stations that uses batteries to capture regenerative braking energy and optimizes ventilation to reduce overall energy consumption and costs. The authors demonstrate that Stochastic Dynamic Programming (SDP) based algorithms outperform Model Predictive Control (MPC) in this application, achieving significant energy and cost savings while maintaining air quality.

What Problem Does This Paper Address?

Subway stations are significant energy consumers, accounting for about a third of a subway system's total energy usage. A major portion of this consumption comes from ventilation systems, which run at high capacity to maintain air quality, especially due to particles generated during train braking. This paper addresses the challenge of reducing this energy consumption and associated costs by proposing an integrated system that recovers braking energy using batteries and intelligently controls ventilation. The core problem is how to optimally manage these energy flows and ventilation in the face of uncertain braking energy availability and demand, while ensuring air quality standards are met.

Methods

The paper proposes an energy management system (EMS) that leverages a battery for regenerative braking energy storage and controls ventilation airflow. The core of the EMS is an optimization problem solved using two main classes of algorithms:

  • Model Predictive Control (MPC): A widely used technique that solves deterministic optimization problems at each time step, using a forecast of future uncertainties.
  • Stochastic Dynamic Programming (SDP): A framework that explicitly accounts for uncertainty. The authors explore two SDP-based algorithms:
  • Offline-online SDPO: This involves an offline backward functional recursion to compute value functions and an online exhaustive search for control decisions.
  • Offline-online SDPA: An extension of SDPO that augments the state to handle dependencies between noises, aiming for potentially better performance.

The system dynamics are modeled using differential equations for battery state of charge and PM10 concentration in the station air, which are then discretized for the optimization algorithms.

Datasets

The paper does not explicitly mention using pre-existing public datasets. Instead, it describes generating:

  • 5,000 "optimization scenarios": These were generated using a calibrated rule based on realistic data for braking energy, demand, train arrivals, and outdoor air quality. These scenarios were used to design and tune the MPC and SDP algorithms.
  • 10,000 "assessment scenarios": These were generated independently from the optimization scenarios and used solely for evaluating the performance of the developed strategies in an "out-of-sample" manner.

Key Insights and Ideas

  • Synergistic Control: The paper highlights the benefit of jointly controlling battery energy storage and ventilation. This integrated approach allows for more efficient energy recovery and consumption reduction than managing these systems independently.
  • SDP's Advantage in Uncertainty: For this specific application with inherent uncertainties in braking energy, SDP-based methods demonstrate superior performance compared to MPC. This is attributed to SDP's ability to explicitly model and react to these uncertainties.
  • Offline vs. Online Computation Trade-offs: While SDP algorithms (especially SDPA) require significant offline computation, their online computation time is very fast, making them suitable for real-time implementation within the subway's control architecture. MPC, on the other hand, can have computationally expensive online optimization steps.
  • Air Quality as a Constraint, Not Just an Outcome: The system aims to maintain air quality while reducing energy. The discomfort associated with poor air quality is explicitly priced into the optimization objective, ensuring that energy savings do not come at the expense of passenger or worker health.

Key Results

  • Significant Savings: Both SDP and MPC algorithms achieved substantial energy and cost savings, on the order of one-third, compared to a baseline scenario without a battery or smart control.
  • SDP Outperforms MPC: The SDP-based algorithms, particularly SDPA, showed a slight but consistent advantage over MPC in terms of economic savings (around 3%) and energy savings.
  • Computational Efficiency: While SDPA has a higher offline computation cost, its online computation time is very low (milliseconds), making it practical for real-time control. MPC's online computation can be more demanding.
  • Air Quality Maintained: The proposed strategies successfully maintained air quality comparable to the reference case, despite reducing ventilation energy consumption.

Limitations and Future Work

  • Battery Cost Not Included: The reported savings do not account for the cost of the battery itself, which would be a crucial factor in a real-world economic analysis.
  • Simplified Models: The models used for battery dynamics, air quality (PM10 concentration), and noise generation are simplified. More complex and accurate models could lead to different results.
  • Deterministic Parameters: Some parameters, such as electricity cost and outside air quality, were assumed to be deterministic. Incorporating uncertainty in these parameters could further refine the optimization.
  • Future Work: The authors suggest completing the analysis with an economic assessment that includes battery costs and practical installation considerations. Further research could explore more advanced optimization techniques and more sophisticated system models.

Summary updated

Abstract

In the Paris subway system, stations represent about one-third of the overall energy consumption. Within stations, ventilation is among the top consuming devices; it is operated at maximum airflow all day long, for air quality reasons. In this paper, we present a concept of energy system that displays comparable air quality while consuming much less energy. The system comprises a battery that makes it possible to recover the trains braking energy, arriving under the form of erratic and strong peaks. We propose an energy management system that, at short time scale, controls energy flows and ventilation airflow. By using proper optimization algorithms, we manage to match supply with demand, while minimizing energy daily costs. For this purpose, we have designed algorithms that take into account the braking variability. They are based on the so-called stochastic dynamic programming (SDP) mathematical framework. We fairly compare SDP-based algorithms with the widespread model predictive control (MPC) ones. First, both SDP and MPC yield energy/money operating savings of the order of one-third, compared to the current management without battery. Second, depending on the specific design, we observe that SDP outperforms MPC by a few percent, with an easier online numerical implementation.
BibTeX
@article{rigaut2018stochastic,
  author = {Tristan Rigaut and Pierre Carpentier and Jean‐Philippe Chancelier and Michel De Lara and Julien Waeytens},
  title = {{Stochastic Optimization of Braking Energy Storage and Ventilation in a Subway Station}},
  journal = {IEEE Transactions on Power Systems},
  year = {2018},
  doi = {10.1109/tpwrs.2018.2873919},
}
Figure