Accurate estimation of battery degradation cost is one of the main barriers for battery participating on the energy arbitrage market. This paper addresses this problem by using a model-free deep reinforcement learning (DRL) method to optimize the battery energy arbitrage considering an accurate battery degradation model. Firstly, the control problem is formulated as a Markov Decision Process (MDP). Then a noisy network based deep reinforcement learning approach is proposed to learn an optimized control policy for storage charging/discharging strategy. To address the uncertainty of electricity price, a hybrid Convolutional Neural Network (CNN) and Long Short Term Memory (LSTM) model is adopted to predict the price for the next day. Finally, the proposed approach is tested on the the historical UK wholesale electricity market prices. The results compared with model based Mixed Integer Linear Programming (MILP) have demonstrated the effectiveness and performance of the proposed framework.

This paper applies advanced battery modeling and multi-objective constrained nonlinear optimization techniques to derive suitable charging patterns for lithium-ion batteries. Three important yet competing charging objectives, including battery health, charging time, and energy conversion efficiency, are taken into account simultaneously. These optimization objectives are first subject to a high-fidelity battery model that is synthesized from recently developed individual electrical, thermal, and aging models. The coupling relationship and multiple timescales among different model dynamics are identified. Furthermore, constraints are considered explicitly on the current, voltage, state-of-charge, and temperature. Such a complex charging problem is solved by using an ensemble multi-objective biogeography-based optimization (EM-BBO) approach. As a result, two charging patterns, namely the constant current-constant voltage (CC-CV) and multistage constant current-constant voltage (MCC-CV), are optimized to balance various combinations of charging objectives. Different trade-offs and sensitive elements are compared and analyzed based on the Pareto frontiers. Illustrative results demonstrate that the proposed strategy can effectively offer feasible health-conscious charging with desirable trade-offs among charging speed and energy conversion efficiency under different demand priorities.

As a large energy consumer, the railway systems in many countries have been electrified gradually for the purposes of performance improvement and emission reduction. With the widespread utilization of energy-saving technologies such as regenerative braking techniques, and in support of the full electrification of railway systems in a wide range of application conditions, energy storage systems (ESSes) have come to play an essential role. In this paper, some recent developments in railway ESSes are reviewed and a comprehensive comparison is presented for various ESS technologies. The foremost functionalities of the railway ESSes are presented together with possible solutions proposed from the academic arena and current practice in the railway industry. In addition, the challenges and future trends of ESSes in the railway industry are briefly discussed.

A modern DC microgrid often comprises renewable energy sources (RESs) such as Photovoltaic (PV) generation units, battery energy storage systems (BESSs), and local load, and it is also connected to the utility grid through a point of common coupling (PCC). While most existing approaches have to rely on communication links to achieve desirable control performance, this paper proposes a novel control strategy without resorting to the communication links. This is achieved by assigning BESSs as master units and regulating the DC bus voltage with a novel state-of-charge (SoC)-based droop control, where the BESSs coordinate the slave units (e.g. RES, utility grid) with the aid of the DC bus signaling (DBS) technique to avoid overcharging and over-discharging of these BESSs. In the proposed droop control, the reference voltage for these BESSs is designed for coordinated operation between BESSs and utility grid, it is maintained constant in normal SoC range, which can reduce DC voltage variation. Droop coefficients designed for SoC balance of BESS are dynamically adjusted based on their own SoC values. Furthermore, the preset maximum deviation between the reference voltage and DC bus voltage ensures reliable coordinated operation. Real-time hardware-in-loop (HIL) tests considering three different scenarios are conducted to validate the effectiveness of the propose method.

A real-time state-of-charge (SOC) estimator based on the signals obtained from a Fibre Bragg Grating (FBG)-based sensor system is reported. The estimator has used a dynamic time-warping algorithm to determine the best fit, employing previously obtained experimental data. The strain data used were obtained from the optical signal monitored, providing the input to a supervised learning algorithm. The results achieved show a good match with those from conventional techniques, achieving a ~2% accuracy with a ~1% SOC resolution. The system has been successfully applied to a ‘proof of concept’ demonstrator, using a battery-operated train, illustrating as a result the way in which the real-time SOC estimator could be employed to enhance safety in the growing electrical vehicle industry.