Imbalanced sample fault diagnosis method for rotating machinery in nuclear power plants based on deep convolutional conditional generative adversarial network

Zhichao Wang,Hong Xia,Jiyu Zhang,Bo Yang,Wenzhe Yin Korean Nuclear Society 2023 Nuclear Engineering and Technology Vol.55 No.6

Rotating machinery is widely applied in important equipment of nuclear power plants (NPPs), such as pumps and valves. The research on intelligent fault diagnosis of rotating machinery is crucial to ensure the safe operation of related equipment in NPPs. However, in practical applications, data-driven fault diagnosis faces the problem of small and imbalanced samples, resulting in low model training efficiency and poor generalization performance. Therefore, a deep convolutional conditional generative adversarial network (DCCGAN) is constructed to mitigate the impact of imbalanced samples on fault diagnosis. First, a conditional generative adversarial model is designed based on convolutional neural networks to effectively augment imbalanced samples. The original sample features can be effectively extracted by the model based on conditional generative adversarial strategy and appropriate number of filters. In addition, high-quality generated samples are ensured through the visualization of model training process and samples features. Then, a deep convolutional neural network (DCNN) is designed to extract features of mixed samples and implement intelligent fault diagnosis. Finally, based on multi-fault experimental data of motor and bearing, the performance of DCCGAN model for data augmentation and intelligent fault diagnosis is verified. The proposed method effectively alleviates the problem of imbalanced samples, and shows its application value in intelligent fault diagnosis of actual NPPs.

Bridge weigh-in-motion through bidirectional Recurrent Neural Network with long short-term memory and attention mechanism

Zhichao Wang,Yang Wang 국제구조공학회 2021 Smart Structures and Systems, An International Jou Vol.27 No.2

In bridge weigh-in-motion (BWIM), dynamic bridge response is measured during traffic and used to identify overloaded vehicles. Most past studies of BWIM use mechanics-based algorithms to estimate axle weights. This research instead investigates deep learning, specifically the recurrent neural network (RNN), toward BWIM. In order to acquire the large data volume to train a RNN network that uses bridge response to estimate axle weights, a finite element bridge model is built through the commercial software package LS-DYNA. To mimic everyday traffic scenarios, tens of thousands of randomized vehicle formations are simulated, with different combinations of vehicle types, spacings, speeds, axle weights, axle distances, etc. Dynamic response from each of the randomized traffic scenarios is recorded for training the RNN. In this paper we propose a 3-stage Bidirectional RNN toward BWIM. Long short-term memory (LSTM) and attention mechanism are embedded in the BRNN to further improve the network performance. Additional test data indicates that the BRNN network achieves high accuracy in estimating axle weights, in comparison with a conventional moving force identification (MFI) method.

Non-salt roasting mechanism of V–Cr slag toward efficient selective extraction of vanadium

Zhichao Yao,Qing Zhang,Ling Wang,Wenjuan Zhang,Baozhong Ma,Chengyan Wang 한국공업화학회 2023 Journal of Industrial and Engineering Chemistry Vol.126 No.-

Developing a clean method for the sustainable utilization of reducing slag containing vanadium and chromium(V–Cr slag) is necessary, with the non-salt roasting method exhibiting considerable potential. Inthis study, we conducted a detailed analysis of the mechanisms driving non-salt roasting of the V–Cr slagby investigating phase transformations, valence transition, and redistribution of vanadium and associatedtransition elements. Based on our findings, we performed efficient selective extraction experimentsand revealed that spinel (Fe,Mn)(V,Cr)2O4, the carrier of vanadium, is converted to acid-soluble products,including VO2, Mg2V2O7, and Mn2V2O7, and acid-insoluble products, such as (Fe,Cr)2O3, TiO2, MnO2, andFe2TiO5 at a roasting temperature of 700 C; At roasting temperatures > 850 C, MnO2 is converted to halfMn2V2O7 and half (Fe,Cr,Mn)2O3. Vanadium is separated from Fe, Cr, Ti, Si, and half Mn via selective leaching. Under optimized conditions, the leaching efficiency of V reached 94.55%, and those of Fe, Cr, Ti, Si,and Mn were 0.71%, 0.50%, 0.44%, 0.23%, and 50%, respectively. This study provides a technical and theoreticalbasis for cleaner and safer utilization of V–Cr slag.

Global Periodic Solutions in a Delayed Predator-Prey System with Holling II Functional Response

Jiang, Zhichao,Wang, Hongtao,Wang, Hongmei Department of Mathematics 2010 Kyungpook mathematical journal Vol.50 No.2

We consider a delayed predator-prey system with Holling II functional response. Firstly, the paper considers the stability and local Hopf bifurcation for a delayed prey-predator model using the basic theorem on zeros of general transcendental function, which was established by Cook etc.. Secondly, special attention is paid to the global existence of periodic solutions bifurcating from Hopf bifurcations. By using a global Hopf bifurcation result due to Wu, we show that the local Hopf bifurcation implies the global Hopf bifurcation after the second critical value of delay. Finally, several numerical simulations supporting the theoretical analysis are given.

MODEL-BASED NH3 ADAPTATION METHOD FOR SCR CONTROL

Tiantian Wang,Zhichao Huang,Xuwei Luo,Jie Hu,Zhi Wang,Yayu Cheng 한국자동차공학회 2022 International journal of automotive technology Vol.23 No.6

Selective Catalytic Reduction (SCR) system is widely used for diesel nitrogen oxide (NOx) control, but the closed-loop control and adaptive correction method which based on NOx sensor signal will be affected by the working characteristics of NOx sensor, such as the cross-sensitivity of ammonia (NH3). Based on the highly linear correlation between conversion efficiency and ammonia to NOx ratio (ANR) at insufficient supply level, a model-based adaptive correction method for urea injection quantity is proposed to eliminate this effect, which can effectively correct the deviation caused by multiple sources when using accurate models. The difference between the conversion efficiency of current SCR system and which generated from model is calculated by stopping injection and under injection actions to obtain an accurate adaptation factor, and the basic injection quantity can be modified by this factor. The maximum error of the 18 calculated correction factors from 18 WHTC cycles is only 1.1 % when using standard AdBlue; After urea solution concentration changed, the calculated correction factor error is within 5 % when comparing with the theoretical correction factor after 2 ~ 3 rounds adaptation function running. The adaptation function keeps the reductant supply to adapt the deviation caused by catalyst aging, reductant concentration drift and supply accuracy variation only with slightly influence of tailpipe NOx emission (less than 2 % in average), which ensure SCR system always stay in appropriate dosing control state, make best for NOx emissions and NH3 emissions control.

Health Care System Based on Wireless Sensor and Mobile Communications

Xiaoli Wang,Zhichao Wang 보안공학연구지원센터 2016 International Journal of Smart Home Vol.10 No.12

With health wearable device becoming increasingly popular on the market, health has become a hot topic of concern. In view of the aging process as the population appears empty nest, elderly people living alone health care issues, a three-tier health care system architecture design based on the “health care acquisition layer embedded terminal - server software system layer-routing layer” is presented in this paper. Aiming for systematic design plan, this paper studied the implementation of key technologies including the following part: the physiological parameter acquisition, ZigBee networking, GPRS mobile communication and Bluetooth Low Energy (BLE) etc. System enables the user to sign of physiological parameters of the real-time dynamic acquisition upload, and through the Android APP and Web two ways to remotely view, providing real-time reference data for the user and medical personnel.

Numerical Simulation of the Creep Behavior of Beishan Deep Granite Tunnel under the Coupling Thermal -Stress Field

Jiawei Wang,Ju Wang,Zhichao Zhou,Peng Wu,Haoran Sun,Jiale Dou,Nan Li,Xianzhe Duan 대한토목학회 2024 KSCE Journal of Civil Engineering Vol.28 No.4

This study conducts a three-dimensional numerical simulation of the creep behavior of deep granite tunnels at 560 meters underground during the coupled thermal-stress process, with an aim to elucidate the effect of the temperature and stress fields on the creep behavior of these deep hard rock tunnels. A 100-h creep period was set, and the mechanical structure of the experimental chamber was accurately replicated at a 1:1 scale, considering the actual mechanical structure of the granite tunnels in the Beishan underground laboratory. The simulation results can demonstrate that: 1) The maximum stress at 90°C and 50°C are 2.86 and 1.91 times than that at 23°C, respectively, demonstrating significant strain accumulation in the deep granite tunnels at the surface. This phenomenon can primarily be attributed to the thermal stress resulted from the coupling between temperature and stress. 2) The maximum creep at 90°C and 50°C is 16 and 3.5 times than that at 23°C. Under the influence of thermal coupling, the creep increases significantly with increasing temperature, indicating that temperature is an important factor influencing creep in granite. 3) Compared with variations in the stress field, the temperature field emerges as the most critical factor influencing granite creep.

Simulation and performance evaluation of the anoxic/anaerobic/aerobic process for biological nutrient removal

Zhen Zhou,Zhichao Wu,Zhiwei Wang,Shujuan Tang,Guowei Gu,Luochun Wang,Yingjun Wang,Zhiling Xin 한국화학공학회 2011 Korean Journal of Chemical Engineering Vol.28 No.5

As a modified configuration of the conventional anaerobic/anoxic/aerobic (AAO) process, a novel anoxic/anaerobic/aerobic (Reversed AAO, RAAO) process has been extensively applied in domestic wastewater treatment plants (WWTP). In this study, the Activated Sludge Model No. 2d (ASM2d) and a secondary clarifier model were calibrated and applied to simulate a pilot-scale RAAO test and evaluate the operational performance of the RAAO process. For calibration of the biological model ASM2d, only four kinetic parameters were adjusted to accurately simulate in-process variations of ammonium, nitrate and phosphate. Simulation results by the calibrated model demonstrated that phosphorus accumulating organisms (PAO) in the RAAO process (0.243 gP·(gCOD)^−1) contains less poly-phosphate than the AAO process (0.266 gP·(gCOD)^−1). With the increasing mixed liquor recirculation ratio in the RAAO process,the fraction of heterotrophic biomass and autotrophic biomass both increased, whereas the PAO decreased owing to adverse effects of electron acceptors on phosphorus release and poly-hydroxy-alkanoates synthesis.

Enhanced morphology and hydrophilicity of PVDF flat membrane with modified CaCO3@SMA additive via thermally induced phase separation method

Zhichao Zhang,Wei Wang,Xin Xu,Xi Liu,Yuanling Li,Peng Zhang 한국공업화학회 2022 Journal of Industrial and Engineering Chemistry Vol.107 No.-

Modified CaCO3@SMA nanoparticles obtained by coordination reaction between poly(styrene-co-maleicanhydride) (SMA) and calcium carbonate (CaCO3) nanoparticles were adopted as additive to preparepolyvinylidene fluoride (PVDF) flat membrane via thermally induced phase separation (TIPS) methodwith dioctyl phthalate (DOP) and dibutyl phthalate (DBP) as mixed diluent. The CaCO3 nanoparticlesmodified by SMA effectively reduced the adverse agglomeration of nanoparticles and made the additivedisperse evenly in PVDF matrix. The presence of modified CaCO3@SMA changed the membrane morphologyfrom dispersed spherulites with large pores to fuzzy dendrite structures with uniform pore sizes. Themembrane pore sizes, pore size distribution and tensile strength were significantly improved comparedto both virgin membranes and those containing unmodified CaCO3 nanoparticles. After the CaCO3 waspickled, the porosity and the connectivity between membrane pores were greatly enhanced, resultingin a significant increase in pure water flux. At the same time, the amphiphilic SMA fixed in and on themembrane surface improved the hydrophilic and anti-fouling properties demonstrated in a three-cycletest. The present study provided a potential TIPS method for the fabrication of PVDF membrane combinedwith a simple strategy of modified inorganic particle additive.

Fabrication and Magnetic Properties of A New Fe-based Amorphous Compound Powder Cores

Wang Xiangyue,Guo Feng,Lu Caowei,Lu Zhichao,Li Deren,Zhou Shaoxiong 한국자기학회 2011 Journal of Magnetics Vol.16 No.3

A new Fe-based amorphous compound powder was prepared from Fe-Si-B amorphous powder by crushing amorphous ribbons as the first magnetic component and Fe-Cr-Mo metallic glassy powder by water atomization as the second magnetic component. Subsequently by adding organic and inorganic binders to the compound powder and cold pressing, the new Fe-based amorphous compound powder cores were fabricated. This new Fe-based amorphous compound powder cores combine the superior DC-Bias properties and the excellent core loss. The core loss of 500 ㎾/㎥ at Bm = 0.1T and f = 100 ㎑ was obtained When the mass ratio of FeSiB/FeCrMo equals 3:2, and meanwhile the DC-bias properties of the new Fe-based amorphous compound powder cores just decreased by 10% compared with that of the FeSiB powder cores. In addition, with the increasing of the content of the FeCrMo metallic glassy powder, the core loss tends to decrease.