Mingda Zhu

Mingda Zhu

Auto-generated profile, based on publicly available information

Mingda Zhu

College of Information Science and Engineering, China University of Petroleum, Beijing, China

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

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

Representative Publications

View All Publications
6 July 2020 in IEEE Access
Verified
Journal Article
Open access
MZMingda Zhu
Mingda Zhu
NLNa Liu
Na Liu
Low-field NMR technology has been widely used in many fields, among which the most representative method is the use of CPMG sequence for T2 measurement. Important parameters such as permeability, saturation and fluid type of porous media can be obtained by T2 spectrum inversion calculation of NMR echo sequence signal. However, T2 spectrum inversion calculation is easily affected by noise signals. Improving the signal-to-noise ratio (SNR) of echo signal is one of the key problems with the application of low-field NMR technology. Therefore, in order to obtain more accurate analysis results, it is necessary to develop corresponding noise reduction methods. The past researches on NMR signal noise reduction mainly focus on echo string. In this paper, a complete NMR signal model is constructed, and the problem of the echo and echo string de-noising is discussed. FIR filter and orthogonal modulation filter (OMF) are studied for noise reduction of echo. To solve the problem of echo string de-noising, an improved complementary ensemble empirical mode decomposition (CEEMD) threshold de-noising method is proposed and compared with the heuristic threshold de-noising method based on sym4 wavelet and the previous empirical mode decomposition (EMD) threshold de-noising method. Based on the construction model and practical verification, it is proved that the proposed method can invert the more accurate T2 spectrum while improving the signal-to-noise ratio. The research results of this paper provide a strong support for the application of low-field NMR technology in the environment of strong noise.

20 September 2020
Verified
Conference Paper
Closed access
JZJunnan Zhang
Junnan Zhang
MZMingda Zhu
Mingda Zhu
LPLei Peng
Lei Peng
Parking data is vulnerably affected by spatiotemporal characteristics and surrounding societal events, causing the latent features of the parking data are hard to learned by GANs if solely given the time-series parking data. Hence it is impossible to generate the desired data with high quality. In this paper, we propose a multi-conditional GAN, named MCGAN to refine the generating process and optimize the generating quality via introducing external customized extendable conditions related to the parking data samples. These conditions, in forms of condition tensors in MCGAN, can help the network learn the features introduced by each defined condition and will reproduce, even combine them in the later generating process, achieving the better result. The experiments show the working process of MCGAN is not different very much from GANs, but the generating quality get improved greatly if given the output expectation more specifically.