Energy management strategy of urban rail hybrid energy storage system based on fuzzy logic system

Energy management is an important link in the effective functioning of hybrid energy storage systems (HESS) within urban rail trains. This factor significantly impacts the operational stability and economic efficiency of urban rail systems. Safety issues arise from DC bus voltage fluctuations due to varying train conditions. To address these issues, this paper proposes an energy management strategy for the urban rail HESS, which builds upon a traditional double closed-loop control strategy. This strategy incorporates a fuzzy logic system to regulate power distribution parameters and discharge thresholds dynamically. Real-time monitoring of the state of charge (SOC) of the energy storage device informs the decision-making process, utilizing a fuzzy inference system to effectively regulate power and energy distribution in response to the distinct characteristics of batteries and super-capacitors. This approach aims to mitigate power fluctuations within the traction network. To validate the effectiveness of the proposed energy management strategy, the study conducts a comprehensive simulation experiment using MATLAB/Simulink. The simulation results confirm the efficacy of the proposed strategy, highlighting its potential to enhance the stability and overall performance of urban rail train operations.


Introduction
Trains often switch traction braking conditions, causing significant voltage fluctuation in the main traction network and affecting train service life.Employing energy storage equipment is crucial for managing network pressure fluctuations and peak braking energy.
Zhao et al. [1] explored the dynamic selection of charge and discharge thresholds, optimizing the energy-saving effect of a single super-capacitor energy storage system (ESS).This approach facilitates more efficient utilization of the ESS.Yang, Luo et al. [2,3] expanded on the traditional control strategy and incorporated a fuzzy inference system to fine-tune the voltage threshold of super-capacitors, resulting in significant voltage stabilization and improved energy efficiency.Similarly, Cheng, Chen et al. [4,5] leveraged a low-pass filtering method for power distribution, allowing different ESS within the HESS to maximize their respective advantages.Tang and Liu [6] employed a second-order high-pass filtering algorithm with variable time constants to manage power distribution in the HESS, thereby preventing overcharging and over-discharging of the energy storage equipment The HESS comprises high-energy-density battery packs and supercapacitor units, meeting high energy and power density needs, respectively.They are frequently interconnected in a specific configuration to establish the HESS [7,8].
Drawing from the insights of the research above, this paper adopts a second-order low-pass filter with a dynamic time constant power allocation strategy combined with a fuzzy logic system for realtime adjustments.Emphasizing the reduction of DC bus voltage fluctuations and addressing battery cycle longevity concerns, the proposed energy management strategy is rigorously validated through MATLAB/SIMULINK simulations.

Analyzing bus voltage fluctuations in traction networks
When the torque generated by the traction motor of the train is transmitted to the rail, the rail will produce an external force on the vehicle, thus pushing the train to run.In this case, the bus voltage drops when the train is accelerating.Conversely, when the train is braking, the traction motor transitions to the generator state, producing torque that opposes the initial traction.The torque produced by the generator during this phase is commonly referred to as braking torque, serving to decelerate the train at a consistent rate and facilitating the effective execution of the braking function.However, due to the limitations of the 24-pulse uncontrolled rectifier, the system cannot feed energy back to the urban power grid.Consequently, excess energy is stored in the support capacitor.Thus, the joint effects of acceleration and braking in urban rail systems fundamentally contribute to fluctuations in bus voltage.

Traction and braking characteristics of urban rail trains
As shown in Figure 1, the traction characteristics of urban rail EMUs in the traction process usually comprise three sections.The low-speed range of the traction process is composed of constant torque (CT) area, and the medium-high-speed area is composed of constant power (CP) area and natural characteristic (NC) area.The braking process is divided into the natural characteristic area and the constant traction characteristic area [9].  Figure 1 illustrates that the constant power zone's characteristic curve is flat, and the range of the constant power zone is wide, which is more suitable for energy storage equipment.However, there is no constant power zone in the braking process, which leads to a peak-shaped characteristic curve and is more suitable for power-type energy storage equipment.Therefore, to optimize the use of different ESSs in hybrid energy storage systems, a comprehensive assessment of the strengths of HESS and their match with different power characteristics is necessary.

Analysis of urban rail power supply system, including HESS
The urban rail transit traction power supply system consists mainly of Figure 2 and various components.Because of the dense distance between stations, trains need to start and stop frequently, which will cause the sudden increase or decrease of large-capacity loads such as traction locomotives and bring about DC bus voltage fluctuation, which will make urban rail vehicles run under the condition of traction network voltage with large voltage fluctuation long period, this situation significantly impacts the long-term stable operation of locomotives.Consequently, the incorporation of an energy storage device becomes imperative.This addition enables the absorption of substantial braking-generated energy during the train's braking phase so that the train can release the stored energy under traction conditions, thus reducing the voltage fluctuation range on the contact network.

HESS for urban rail trains
The main substation uses a step-down transformer to reduce the high voltage to medium voltage.Then, it sends it to the traction substation for step-down rectification to 1500 V/750 V DC power, which is then transmitted to the traction DC bus for the train motor.The HESS primarily comprises a bidirectional DC/DC converter (BDC) and an energy storage device [10].As depicted in Figure 3, a non-isolated bidirectional DC/DC charging and discharging device can achieve bidirectional power flow by changing the current direction.

HESS control strategy
This paper proposes a novel energy management approach for the urban rail HESS based on the conventional double closed-loop control strategy.The approach integrates a fuzzy logic system to regulate power distribution parameters and discharge thresholds.The control strategy for the HESS is segmented into four components, as illustrated in Figure 4. Initially, the State of Charge (SOC) for both the super-capacitor and the SOC of the lithium-ion battery serve as inputs for the fuzzy reasoning module.Subsequently, the fuzzy logic reasoning module dynamically adjusts HESS's discharge threshold and the power distribution module's filtering time constant in real time.

Energy management of energy storage system under power distribution
In energy management, employing a second-order low-pass filter (LPF) is crucial for decomposing the ESS's target power.To separate the high and low-frequency signals, we use the high-power density characteristics of super-capacitor energy storage devices and the high-energy density characteristics of battery energy storage devices.At the same time, a fuzzy controller dynamically regulates the filter's time constant to establish the allocated reference power for each component.

Fuzzy controller design
A comprehensive adjustment of the overall reference power is implemented to enhance the adaptability of the battery and super-capacitor energy storage device across various target powers and charging operational states, while acknowledging the distinctions between energy-type and power-type storage devices.This involves the dynamic fine-tuning charging and discharging thresholds and the filtering time constant using fuzzy control.The inference system has two inputs and two outputs.The input

Power distribution module.
In the control strategy shown in Figure 4, the reference power Phess_ref(s) is decomposed by a second-order low-pass filter into two components: low-frequency battery reference power Pbat_ref(s) and high-frequency super-capacitor reference power Psc_ref(s), where the decomposed relationship is as follows: The operator "s" represents a differential operator, while "T" signifies the low-pass filtering time constant.The time constant, derived from the fuzzy logic reasoning system, relies on the state of charge for each ESS.Consequently, the filter time constant "T" directly impacts the filter's cut-off frequency, influencing the decomposed battery reference power.Notably, a higher filter time constant "T" leads to a reduced cut-off frequency, resulting in a narrower frequency range and diminished power that the battery can compensate for.Simultaneously, this adjustment elevates the output power of the supercapacitor, thereby alleviating the burden on the battery and extending its overall service life.

Energy storage system threshold module.
By comparing the discharge threshold obtained by the fuzzy controller with DC bus voltage, three different working modes of the energy storage system are obtained, as shown in Figure 7.These modes are employed to prevent any potential mishaps in the energy storage system.

Simulation and result analysis
Based on the fuzzy logic system, a simulation model is developed using MATLAB/Simulink to validate the suggested dynamic threshold energy management control approach for urban rail HESS.The comprehensive simulation encompasses various working conditions, such as acceleration, idle running, and braking.The specific parameters employed for the simulation are detailed in Table 1.As depicted in Figure 8, during the train's traction phase (0.2 s to 0.8 s), if the DC network voltage exceeds the acceptable range, the HESS's discharge threshold is triggered to supply energy.Without the use of an energy storage device, the DC traction network experiences severe voltage fluctuations during the idle phase (0.8 s to 1.4 s) and the braking phase (1.4 s to 2 s).These fluctuations cause the braking resistor to consume energy.
However, with the incorporation of the ESS, the fluctuations in the DC bus voltage are mitigated, enabling the capture of excess braking energy for subsequent supply during the train's traction phase.Furthermore, the simulation results affirm the efficacy of the voltage threshold set by the fuzzy logic system, maintaining the DC traction network voltage within the desired range of 1580 V to 1750 V.As shown in the state of charge distribution diagram of Figure 9 and the power distribution diagram of Figure 10, when the train is in traction condition (0.2-0.8 s) and reaches the voltage threshold of energy storage system discharge, the super-capacitor first starts to discharge rapidly, while the battery energy storage system slowly starts to discharge after receiving the low-frequency power command due to the existence of power distribution link; 0.6-1.1 s super-capacitor energy storage system avoids overdischarge, and restores its charged state through internal circulation of energy storage system.Subsequently, during the braking phase (1.4 s to 2 s), the super-capacitor energy storage system swiftly absorbs a significant amount of regenerative braking energy.At 1.5 s, the system initiates internal circulation to avert the overcharging of the super-capacitor ESS, thereby validating the efficacy of the power distribution link and its capacity to allocate power efficiently.

Conclusion
This article proposes an energy management control strategy based on fuzzy controllers for urban rail power supply systems, including HESS.A MATLAB/Simulink simulation model is constructed to validate the proposed strategy.The simulation results affirm that the suggested approach ensures the stability of the DC traction network voltage and adeptly regulates the overall discharge of the ESS in line with the specific operational conditions of urban rail trains and the distinctive strengths of individual energy storage devices.
Moreover, the proposed strategy effectively mitigates issues associated with overcharging and overdischarging of the ESS while simultaneously attenuating the adverse effects of rapidly fluctuating power on the battery.As a result, this strategy serves to extend the service life of the battery ESS, signifying its substantial potential for enhancing the overall operational efficiency of urban rail transit systems.

Figure 2 .
Figure 2. Urban rail power supply system, including HESS.
signals are SOCsc of the super-capacitor and SOCbat of the battery, shown in the input membership function in Figure5The discharge threshold reference value and filter time constant 'T' are output by a fuzzy inference system, with its output membership function displayed in Figure6.Because the output universe differs from the universe of corresponding fuzzy sets, the inputs VL, M, VH, S, and L are double Gaussian functions.L and H are Gaussian functions for fuzzification, and the outputs are fuzzified by membership functions determined by triangle functions.

Figure 9 .
Figure 9.Comparison of residual state of charge between battery and superelectric power.

Figure 10 .
Figure 10.Power comparison between battery and super-capacitor.