Cheng Gong

Cheng Gong

Auto-generated profile, based on publicly available information

Cheng Gong

School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, USA

Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
26 April 2021 in IEEE Transactions on Transportation Electrification
Verified
Journal Article
Closed access
JLJinlin Liao
Jinlin Liao
GYGuang Yang
Guang Yang
SZSw Zhang
Sw Zhang
FZFeng Zhang
Feng Zhang
CGCheng Gong
Cheng Gong
At metro stations with very large passenger volumes and severe congestion, random disturbances often occur, resulting in the original offline optimal timetable no longer applicable. Deep reinforcement learning has the advantages of self-learning and online learning, making it possible to solve the energy-aimed train timetable rescheduling (ETTR) problem under random disturbances. In this article, a deep reinforcement learning approach (DRLA) has been proposed and applied to the ETTR problem to reschedule the train timetable in order to achieve the online optimal timetable for the minimum energy usage under disturbances. The proposed DRLA has the advantages of a real-time performance, high learning efficiency, and better energy-saving effect compared with the traditional heuristic algorithms such as genetic algorithm (GA), deep learning methods such as the combination of improved GA and long short-term memory (IGSA-LSTM), or other reinforcement learning algorithms such as deep deterministic policy gradient (DDPG) algorithm. Various experiments are conducted to compare the performance of DRLA and other algorithms. The experimental results indicate that in the two-train metro system and the five-train metro system, DRLA can save an average of energy by 5.11% and 7.29% with an average reaction time of only 0.0013 and 0.0019 s against disturbances, respectively.

24 June 2021 in IEEE Transactions on Industrial Electronics
Verified
Journal Article
Closed access
CGCheng Gong
Cheng Gong
FDFang Deng
Fang Deng
In recent years, Vernier permanent magnet machines (VPMs) have gained much interest in academia. However, due to the large leakage flux, the power factor of Vernier machines is too low to be applied in real industry. In this article, a novel split-tooth, concentrated-winding Vernier machine using ferrite magnets is proposed for high-torque, low-speed direct-drive applications in home appliance industry. It uses consequent-pole magnets in the rotor, as well as additional assisted magnets in the stator teeth, in order to reduce the leakage flux and boost the torque density. A systematic way of reducing the torque ripple from 40% to 4% is proposed by carefully choosing the right combination of stator/rotor pole-pair numbers. A comprehensive design optimization is conducted to achieve the optimal performance. Finally, the ac loading is carefully reduced to meet the torque, saturation, power factor and space requirements. The test results of a prototyped machine show that the proposed design can achieve 34% more torque and 20% less loss with a reasonable power factor of 0.63, compared with an optimized PM synchronous machine under the same volume. It is found that the significant torque boost is attributed to the multiple working harmonics of the air-gap flux density.

13 January 2020 in IEEE Transactions on Industry Applications
Verified
Journal Article
Closed access
CGCheng Gong
Cheng Gong
SLSufei Li
Sufei Li
THThomas Habetler
Thomas Habetler

7 September 2019 in Energies
Verified
Journal Article
Open access
PDF file
GYGuang Yang
Guang Yang
FZFeng Zhang
Feng Zhang
CGCheng Gong
Cheng Gong
SZSw Zhang
Sw Zhang
Reinforcement learning has potential in the area of intelligent transportation due to its generality and real-time feature. The Q-learning algorithm, which is an early proposed algorithm, has its own merits to solve the train timetable rescheduling (TTR) problem. However, it has shortage in two aspects: Dimensional limits of action and a slow convergence rate. In this paper, a deep deterministic policy gradient (DDPG) algorithm is applied to solve the energy-aimed train timetable rescheduling (ETTR) problem. This algorithm belongs to reinforcement learning, which fulfills real-time requirements of the ETTR problem, and has adaptability on random disturbances. Superior to the Q-learning, DDPG has a continuous state space and action space. After enough training, the learning agent based on DDPG takes proper action by adjusting the cruising speed and the dwelling time continuously for each train in a metro network when random disturbances happen. Although training needs an iteration for thousands of episodes, the policy decision during each testing episode takes a very short time. Models for the metro network, based on a real case of the Shanghai Metro Line 1, are established as a training and testing environment. To validate the energy-saving effect and the real-time feature of the proposed algorithm, four experiments are designed and conducted. Compared with the no action strategy, results show that the proposed algorithm has real-time performance, and saves a significant percentage of energy under random disturbances.

1 October 2020 in IEEE Transactions on Industry Applications
Verified
Journal Article
Closed access
CGCheng Gong
Cheng Gong
SLSufei Li
Sufei Li
THThomas Habetler
Thomas Habetler
PZPing Zhou
Ping Zhou
High-speed switched reluctance machines (SRMs) have been a timely topic recently. However, for ultra-high-speed SRMs over 100,000 rpm, the acoustic noise is a significant issue in many applications such as high speed spindles, compressors/ turbochargers, flywheels and etc. In this paper, a multi-physics acoustic modeling of ultra-high-speed SRMs is presented based on finite element analysis (FEA). First, a 3D electromagnetic FEA is conducted to calculate the radial force on the stator teeth. Then, a modal analysis and a harmonic response analysis are shown in the frequency domain to estimate the natural frequencies and vibration of the whole motor. An acoustic response analysis is done to estimate the sound pressure level (SPL) using the superposition of the surface vibration. To verify the model, experiments are done to measure and compare the natural frequencies and the SPL of a 4/2 ultra-high-speed SRM at 100,000 rpm. The results show that the SPL of the motor is as high as 115 dB at 10 mm away from the motor, which matches the simulation result (113 dB) well. The impact of the bearing is proven to be negligible compared with the noise generated by the radial force. Finally, a parametric study example of using the proposed model to guide the design of SRMs has been conducted. It is shown that "hot dog" shaped motors have less noise than "pancake" shaped motors under the same motor volume, speed and power.

30 August 2018 in IEEE Transactions on Industry Applications
Verified
Journal Article
Closed access
CGCheng Gong
Cheng Gong
SLSufei Li
Sufei Li
THThomas Habetler
Thomas Habetler
JRJose Alex Restrepo
Jose Alex Restrepo
BSBrian Soderholm
Brian Soderholm
High speed switched reluctance machines (SRMs) are gaining popularity in both academia and industrial communities. One of the reasons that limit the maximum speed of SRMs is the speed limit of commercial rotary encoders due to mechanical considerations. In this paper, a novel direct position control for ultra-high speed SRMs is proposed based on non-intrusive reflective sensors. It uses reflective optical sensors to detect the relative position between the rotor and the stator so as to control the on and off states of the switches directly through the analog output trigger signals generated by the optical sensors. In order to generate the maximum output torque, the switching-on and -off angles are optimized based on the equivalent Simulink model with the torque and flux linkage characteristics obtained from a corresponding finite element analysis (FEA) of a high speed SRM. Finally, the proposed method is validated by experiments with a 500 W 4/2 SRM running up to 100,000 rpm. The proposed control system has many advantages such as low cost and virtually no speed limit, which makes it suitable for ultra-high speed applications.