Top Read Articles

    Published in last 1 year |  In last 2 years |  In last 3 years |  All
    Please wait a minute...
    For Selected: Toggle Thumbnails
    Arc modeling and practical simulation application based on ATP-EMTP
    DOU Tingting, YAO Yuanxi, CHEN Peng, LU Deng
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2018, 48 (4): 102-108.   DOI: 10.6040/j.issn.1672-3961.0.2018.129
    Abstract3342)   HTML    PDF(pc) (1438KB)(336)       Save
    The HVDC line and earth electrode line might be constructed using common tower, but there was lack of research on the back flashover performance of HVDC line, whats more, the earth electrode line arcing horn discharge characteristics was also scanty, in the case of tower struck by lightning. The frequency characteristic simulation transmission line, multi-impedance tower model and air gap breakdown model were established by ATP-EMTP, the influence of electrode line on back flashover level for HVDC line was studied, and the arcing horn discharge characteristics was researched. The results showed that, back flashover performance of HVDC line could obviously be improved by earth electrode line, and lasting arc could be established by DC current in the single pole operation mode.
    Reference | Related Articles | Metrics
    A new method for muti-objects image segmentation based on faster region proposal networks
    HUANG Jinchao
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2018, 48 (4): 20-26.   DOI: 10.6040/j.issn.1672-3961.0.2017.592
    Abstract3292)   HTML    PDF(pc) (6800KB)(986)       Save
    Aim at the shortage of a large amount of redundancy and overlaps of conventional semantic segmentation algorithm, these shortages caused the image segmentation results getting lower accuracy and robust. A new algorithm of multi-objects image segmentation based on faster region proposal networks was proposed. A selective search algorithm was used to get the initial proposal boxes; a faster region proposal network was used to get initial image segmentation boxes. In order to validate our proposed algorithm, the VGG16 models that pre-trained on ImageNet was used on this problem. By using COCO dataset and Cityscapes dataset, the model was well fine-tuned. The test dataset was used for testing semantic segmentation and image segmentation. Compared with YOLO algorithm, the experimental results showed that our proposed algorithm increased mAP of 2.16% and 1.55%. The initial image segmentation boxes by faster region proposal networks were best fitted by GrabCut, multi-objects segmentation results were more accurate and robust. Our proposed algorithm got higher accuracy by sacrifice little time consumption, which got more application scenes in multi-object patterns recognition.
    Reference | Related Articles | Metrics
    Smooth walk-to-trot gait transition algorithm for quadruped robot
    XIN Yaxian, LI Yibin, LI Bin, RONG Xuewen
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2018, 48 (4): 42-49.   DOI: 10.6040/j.issn.1672-3961.0.2017.364
    Abstract3280)   HTML    PDF(pc) (4590KB)(783)       Save
    In order to improve the adaptability of quadruped robot for various terrains, different cases that could occur when robot changed its gaits from walk to trot were analyzed and the optimal transition algorithms which could let gait transition more smooth and waste least time when robot kept stability were proposed. In order to ensure the smoothness of transition sequence, the speed formula of time was given to keep the center of gravity constant acceleration. An algorithm named modified wide stability margin method(MWSM)was proposed to offset the stand back influenced by inertial force and caused by the acceleration through adjust the relative position of the trunk and four feet. The model of quadruped robot was constructed based on the robot simulator Webots, and simulation results showed the validity and effectiveness of the algorithm. The approach could be applied in six points of one static walk circle and switched to trot smoothly and steadily.
    Reference | Related Articles | Metrics
    Immune clonal evolutionary algorithm of dynamic economic dispatch considering gas pollution emission
    QIAN Shuqu, WU Huihong, XU Guofeng, JIN Jingliang
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2018, 48 (4): 1-9.   DOI: 10.6040/j.issn.1672-3961.0.2017.369
    Abstract3209)   HTML    PDF(pc) (881KB)(587)       Save
    An immune clonal evolutionary algorithm(ICEA)was proposed by combining the clone selection principle of immune system and the evolution mechanism of genetic algorithm. A kind of dynamic immune selection strategy was introduced and a self-adaption non-uniform mutation operator was proposed. In order to make it suitable for solving dynamic emission economic dispatch(DEED)problem with many constrains, different repair strategies were introduced for the equality and inequality constrains of DEED model. In numerical experiments, ICEAs performance on 10-units system was tested, and several peer algorithms were compared. The simulation results indicated that ICEA had good convergence and global optimization efficiency. The uniformity and ductility of the Pareto optimal frontier obtained by ICEA was better than that of comparison algorithms. The Pareto optimal frontier could provide a more efficient scheduling decision-making approach for power system dispatcher.
    Reference | Related Articles | Metrics
    Tests and comparison of the dehumidification effectiveness of drying techniques involving in atmospheric sampling
    ZHAO Yanan, WANG Xinfeng, LI Rui, CHEN Tianshu, XUE Likun, WANG Wenxing
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2018, 48 (4): 128-136.   DOI: 10.6040/j.issn.1672-3961.0.2018.103
    Abstract2768)   HTML    PDF(pc) (1516KB)(492)       Save
    High humidity and fog water can affect the measurement results of atmospheric pollutants, or even damages the instruments. Therefore, it was necessary to dehumidify the sample air with specific drying techniques when sampling. This study tested and compared the dehumidification effectiveness and the influencing factors of four common drying devices, i.e., cyclone cutter, heating belt, Nafion drying tube, and silicone drying tube. The results showed that the cyclone cutter could effectively remove liquid water and it was suitable trace gases and fine particulate matters. The heating belt quickly reduced the relative humidity of the sample air; however, the relatively humidity exhibited large fluctuation. The average dehumidification efficiency was about 20%~40%. The heating belt was mainly suitable for the thermally stable pollutants. The dehumidification efficiency of Nafion drying tube was usually less than 20%, but it was suitable for all kinds of trace gases and particulate matters. The silica gel drying tube could quickly dry the sample air with stable and high efficiency of about 50% and it was mainly suitable for particulate matters. The use of the above drying techniques would cause a loss of trace gases in certain degree. Among them, the loss caused by heating was highest, about 10%. The loss caused by cyclone cutters and Nafion drying tubes was no more than 10% and even less than 5%. These drying devices had different applicability, advantages, and disadvantages, so it was necessary to take account into the measured component.
    Reference | Related Articles | Metrics
    Image enhancement algorithm based on six dimensional feedforward neural network model
    ZHANG Xianhong, ZHANG Chunrui
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2018, 48 (4): 10-19.   DOI: 10.6040/j.issn.1672-3961.0.2018.063
    Abstract2756)   HTML    PDF(pc) (11598KB)(499)       Save
    Aiming at the problems of weakening the edges caused by filtering denoising, partially indistinct images and low contrast, an image enhancement algorithm based on the six dimension feedforward neural network model was proposed on the basis of fully analyzing the dynamic properties of the model. The experiment showed that the image enhancement algorithm based on the six dimensional feedforward neural network model could better achieve a very good enhancement effect. Compared with other enhancement algorithms, the enhancement effect was clearer and the algorithm was better.
    Reference | Related Articles | Metrics
    Effect of nitrogen species on nitrogen removal performance of Spirulina platensis
    LIU Ting, JIANG Li, ZHOU Weizhi
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2018, 48 (4): 116-121.   DOI: 10.6040/j.issn.1672-3961.0.2018.092
    Abstract2723)   HTML    PDF(pc) (974KB)(195)       Save
    The effects of different nitrogen species(ammonium, nitrite, nitrate, urea)in synthtic secondary effluent on Spirulina platensis growth and nitrogen removal performance was explored. The results showed that Spirulina platensis could grow and effectively remove nitrogen in secondary effluent. After 5 d treatment, concentration of total nitrogen(TN)were decreased to 3.5, 6.8, 4.6, and 4.0 mg/L, respectively, when the nitrogen species were ammonium(29.3 mg/L), nitrite(28.3 mg/L), nitrate(28.6 mg/L)and urea(29.1 mg/L). More than 75% of TN was assimilated in microalgae biomass, and only 4%~10% of TN was removed as gaseous nitrogen. This study provided a theoretical basis for further treatment of nitrogen in secondary effluent, and meanwhile, offered a new solution for microalgae cultivation.
    Reference | Related Articles | Metrics
    Integration transform of dust removal system based on fieldbus compatible technology
    WANG Qi, SUN Zhumei, LIU Shaohong, BAI Jianyun
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2018, 48 (4): 37-41.   DOI: 10.6040/j.issn.1672-3961.0.2018.121
    Abstract2703)   HTML    PDF(pc) (3151KB)(314)       Save
    With the help of a network protocol converter, the Profibus-DP and DeviceNet fieldbus networks communication were realized, which were two networks with heterogeneous protocol. A double levels supervisory control system was built, the Siemens PCS7 as upper main control system, Omron PLC as the control subsystem. By Siemens PCS7 system configuration, the upper level monitoring program CFC configuration, remote monitoring figuration configuration, the gateway PD-100S settings and Omron PLC system configuration, the dust removal system was included in power plant main control system PCS7. The retrofit scheme needed small amount of new equipment. Related hardware and software configuration and parameters setting were simplfy. Under the existing dedusting system data update quantity, the system could reliably realize the remote monitoring. The scheme could provide a reference for the integration of heterogeneous fieldbus system.
    Reference | Related Articles | Metrics
    Fault diagnosis of electromagnetic coil in active magnetic bearing based on current characteristics
    CHENG Xin, ZHANG Lin, HU Yefa, CHEN Qiang, LIANG Dian
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2018, 48 (4): 94-101.   DOI: 10.6040/j.issn.1672-3961.0.2017.617
    Abstract2697)   HTML    PDF(pc) (5624KB)(243)       Save
    Magnetic bearings had been widely used for its excellent performance,but the fault of electromagnetic coil might cause the rotor to lose control and cause serious consequences. A on-line fault diagnosis method of electromagnetic coil formagnetic bearings was proposed through the establishment of mathematical model of digital switching power amplifier output current, the current output characteristic was obtained, and the influence of the coil fault on the output current variation rate was theoretically analyzed. A Matlab/Simulink model was established under the two state modulation,the simulation results demonstrated the theoretical feasibility of the method. A fault diagnosis scheme and related algorithm based on current oversampling was designed.Then a digital switching power amplifier test platform based on digital signal processor(DSP)was built, and the on-line detection experiment of related coil faults was carried out. The experimental results demonstrated the effectiveness of the method discussed in this paper.
    Reference | Related Articles | Metrics
    Fractional internal model control of PWM rectifier based on inverted decoupling
    LI Xiangyu, ZHAO Zhicheng, WANG Wenyu
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2018, 48 (4): 109-115.   DOI: 10.6040/j.issn.1672-3961.0.2017.441
    Abstract2678)   HTML    PDF(pc) (1611KB)(389)       Save
    A novel double closed loop control strategy was proposed for three-phase voltage source PWM rectifier. Based on the mathematical model of PWM rectifier in synchronous rotating coordinate system, the inverted decoupling method was used to realize the complete decoupling of the current loop, and the complicated matrix inversion operation was avoided. According to the principle of internal model control(IMC), the IMC-PI controllers were designed with only one tunable parameter in the inner loop. In the voltage outer loop, a fractional order IMC controller was designed by combining IMC with fractional order control(FOC)method, and the controller parameters were obtained by the cut-off frequency and maximum sensitivity index of the system. The simulation results showed that the proposed method could provide better dynamic performance and disturbance rejection property.
    Reference | Related Articles | Metrics
    Chaos synchronization of a class of fractional-order coronary artery systems
    MENG Xiaoling, WANG Jianjun
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2018, 48 (4): 55-60.   DOI: 10.6040/j.issn.1672-3961.0.2016.463
    Abstract2671)   HTML    PDF(pc) (499KB)(257)       Save
    The problem of chaos synchronization for a class of fractional-order coronary artery systems was studied based on Lyapunov stability theory and fractional-order calculus. Three sufficient conditions were arrived that the fractional order systems was chaos synchronized under appropriate controller. The research conclusion illustrated that systems was chaos synchronization under proper conditions.
    Reference | Related Articles | Metrics
    Image analysis and engineering application of ground penetrating radar in tunnel lining detection
    ZHOU Chenying
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2018, 48 (4): 61-68.   DOI: 10.6040/j.issn.1672-3961.0.2017.015
    Abstract2669)   HTML    PDF(pc) (18207KB)(297)       Save
    A practical and efficient interpreting method of GPR signals was proposed in order to improve the accurate identification of the tunnel lining disease based on the working principle of GPR and practical engineering application experience.The interpretation principles, asthe lining thickness, the distribution of reinforced concrete and steel arch, surrounding rock water rich fissures, karst caves, cavity, non-compaction and radar detachment,were systematically analyzed. A method of high efficient ground penetrating radar signal interpretation for tunnel lining disease detection was proposed. This interpreting method had been analyzed and verified with practical engineering. The reinforcement in the image showed crescent shaped, cavity showed arc reflection in the image, non-compaction performances clutter waveform, the antenna was a strip of bright reflection. The interpreting method provides reference and basis for tunnel lining detection which plays an important role in treating tunnel diseases and evaluating tunnel lining quality.
    Reference | Related Articles | Metrics
    An image feature extraction method based on PDEs
    JIANG Shanshan, YANG Jing, FAN Liya
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2018, 48 (4): 27-36.   DOI: 10.6040/j.issn.1672-3961.0.2018.072
    Abstract2666)   HTML    PDF(pc) (11751KB)(353)       Save
    Further research was conducted on image feature extraction method based on partial differential equations(PDEs). The effect of evolution times on quality of feature, and the reflection of compression function on the quality of feature were studied. Experiment results indicated that the evolution of PDEs could reduce the impact of occlusion and be robust to dark light, but the qualities of the image features could be seriously affected by evolution times of PDEs and compression function and compression speed.
    Reference | Related Articles | Metrics
    Role of siderophore produced by Pseudoaltermonas sp. SCSE709-6 in the removal of Cd2+
    WANG Zhende, HUANG Zhaosong, JIANG Li, ZHOU Weizhi
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2018, 48 (4): 122-127.   DOI: 10.6040/j.issn.1672-3961.0.2018.091
    Abstract2623)   HTML    PDF(pc) (1488KB)(278)       Save
    In order to explore the role of siderophore in the process of removal of heavy metals, Pseudoaltermonas sp. SCSE709-6(P. sp. SCSE709-6), was selected as a representative bacterium. The effect of Cd2+ addition on siderophore production by P. sp. SCSE709-6 and the effect of siderophore addition on the removal of Cd2+ were studied, respectively. Upon O-CAS assay, P. sp. SCSE709-6 showed a vivid positive result for siderophore production. Assays based on chemical properties indicated that siderophore produced by P. sp. SCSE709-6 was carboxylate type. The siderophore yield was correlated with the biomass. Siderophore production was increased first and then decreased when the Cd2+ concentration was 0~0.4 mmol/L, notable this value reached maximum at 0.2 mmol/L. It accelerated the adaptation of P. sp. SCSE709-6 to Cd2+ while siderophore was added to the culture medium, as siderophore could be combined with Cd2+ to reduce the toxicity of cadmium, leading to high removal efficiency of Cd2+. The results provided a scientific explanation of why P. sp. SCSE709-6 is highly efficient in removal of cadmium and P. sp. SCSE709-6 could be recommended as a potential candidate for application in bioremediation of heavy metals.
    Reference | Related Articles | Metrics
    Design and experiment of cascade PID control for Maglev actuator
    WU Huachun, XIE Siyuan, CHEN Changhao
    Journal of Shandong University(Engineering Science)    2018, 48 (4): 88-93.   DOI: 10.6040/j.issn.1672-3961.0.2017.616
    Abstract2615)   HTML    PDF(pc) (9324KB)(223)       Save
    In order to improve the stability of the active vibration isolation system and improve the control quality of the system, the mathematical model of the control channel of the Maglev actuator was obtained by the system identification and the corresponding cascade PID controller was designed. The inner loop of the system was the acceleration loop, which mainly controlled the acceleration of the vibration isolator. The main loop of the system was the position loop, which made the isolator back to the center position and adjusted the reference value of the acceleration controller. The dynamic performance of cascade PID control was theoretically analyzed by the first order system theory; the theoretical results were verified by the establishment of a cascade PID simulation control model in Matlab. Then cascade PID control system was designed by the single degree of freedom experimental platform; the experiment and simulation of acceleration transfer rate was compared; the effect of acceleration control was analyzed. The experimental results showed that the cascade PID control method could make the attenuation range of the control object in the range of 1~25 Hz frequency between -22.5 dB and -2.2 dB, achieved effective vibration control.
    Reference | Related Articles | Metrics
    Dynamics charactersitics of flexible beams undergoing time varying mass
    MA Chicheng, GUO Zonghe, LIU Canchang, DAI Xiangjun, ZHANG Xinong, MAO Boyong
    Journal of Shandong University(Engineering Science)    2018, 48 (4): 78-87.   DOI: 10.6040/j.issn.1672-3961.0.2017.373
    Abstract2603)   HTML    PDF(pc) (2888KB)(240)    PDF(mobile) (9324KB)(4)    Save
    In terms of an elastic beam with a time varying mass, the oscillation equations were derived using mode superposition method, and the influences of the nonstructural damping induced by the changing mass were investigated. The differential equations were solved by self-adaptive Newmark method, then a relative confirmatory experiment was designed, while the change of the mass was realized by controlling the flow of water. The vibration signals were processed using time-frequency analysis toolkit, which showed more features of the time varying mass system in the time-frequency domain. The comparison of experimental results and numerical results demonstrated the feasibility of the proposed approach and the experimental test. The study showed that the nonstructural negative damping induced by the decreasing mass affected the motions significantly, which could not be neglected in the dynamic design of high precision structures as large-scale flexible robotic manipulators.
    Reference | Related Articles | Metrics
    Ratio integral sliding mode synchronization control of entanglement chaotic systems
    MAO Beixing
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2018, 48 (4): 50-54.   DOI: 10.6040/j.issn.1672-3961.0.2017.553
    Abstract2588)   HTML    PDF(pc) (1550KB)(269)       Save
    The problem of sliding mode and ratio integral sliding mode synchronization of a class of entanglement chaotic systems were studied based on sliding mode control in the paper. The surfaces and controllers were designed using sliding mode and ratio integral sliding mode approach. And sliding mode uniform speed reaching law was adopted.Two cases for system trajectory on sliding mode surface and not on sliding mode surface were analyzed based on Lyapunov stability theory.The systems errors could approach to coordinate zero under the corporate action of surfaces and controllers. Two sufficient conditions were arrived for entanglement chaotic systems acquire sliding mode synchronization and integral sliding mode synchronization.The research conclusion illustrated that the master-slave systems of entanglement chaotic systems were sliding mode and ratio integral sliding mode synchronization if proper controllers and sliding mode surfaces was chosen.
    Reference | Related Articles | Metrics
    The influence of soil-cement pile deterioration and modification in salt-water area on road composite foundation
    WANG Zhongxiao, CUI Xinzhuang, CUI Sheqiang, ZHANG Lei, CHE Huaqiao, SU Junwei
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2018, 48 (4): 69-77.   DOI: 10.6040/j.issn.1672-3961.0.2017.580
    Abstract2587)   HTML    PDF(pc) (6447KB)(227)       Save
    In order to reveal the influence of soil-cement pile deterioration and modification in salt-water area on road composite foundation, the physical and chemical properties of soil and water in the Yellow River Delta were analyzed through laboratory experiments, and the mechanical properties of salty soil-cement were studied. Based on Mohr-Coulomb constitutive model and strength reduction method, the influence of soil-cement pile deterioration and modification on the settlement and load-bearing properties of composite foundation were stimulated by FLAC3D. Research showed that, groundwater was basically saltwater with a salinity greater than 5 g/L and the salt was mainly chloride. The corrosion of salt caused deterioration of soil-cement with reduced strength. The deterioration of soil-cement pile increased the settlement of composite foundation and reduced the bearing capacity of pile. It could be effectively improved the bearing capacity of pile and reduced the settlement of composite foundation by using slag powder to modify soil-cement pile. Therefore, in the design of composite foundation, fully considering the deterioration effect of soil-cement pile is the key to ensure that the subgrade has sufficient strength and stability throughout its life cycle.
    Reference | Related Articles | Metrics
    A survey of image captioning methods based on deep learning
    Zhifu CHANG,Fengyu ZHOU,Yugang WANG,Dongdong SHEN,Yang ZHAO
    Journal of Shandong University(Engineering Science)    2019, 49 (6): 25-35.   DOI: 10.6040/j.issn.1672-3961.0.2019.244
    Abstract2298)   HTML63)    PDF(pc) (7881KB)(827)       Save

    Image captioning is the cross-research direction of computer vision and natural language processing. This paper aimsed to summarize the deep learning methods in the field of image captioning. Imgage captioning methods based on deep learning was summarized into five categories: multimodal space based method, multi-region based method, enconder-deconder based method, reinforcement learning based method, and generative adversarial networks based method.The datasets and evaluation metrics were demonstrated, and experimental result of different methods were compared. The three key problems and future research direction for image captioning were presented and summarized.

    Table and Figures | Reference | Related Articles | Metrics
    Object detection of 3D point clouds based on F-PointNet
    Peng WAN
    Journal of Shandong University(Engineering Science)    2019, 49 (5): 98-104.   DOI: 10.6040/j.issn.1672-3961.0.2018.348
    Abstract2243)   HTML40)    PDF(pc) (3128KB)(808)       Save

    Aiming at the problem of poor detection accuracy of the current 3D point cloud object detection model, the F-PointNet model, which directly processed point cloud data, was used to detect cars, pedestrians and cyclists, and the model was fine-tuned to further improve the object detection accuracy. The model was tested by different parameter initialization methods, $\ell $2 regularization and modifying convolution kernels. The experimental results showed that the Xavier parameter initialization method converged faster 0.09s than the truncated normal distribution method, and the vehicle detection accuracy and the cyclists detection accuracy was about 3% and 2% higher respectively. By adding $\ell $2 regularization, the detection accuracy of pedestrians and cyclists was increased by about 2% and 1% respectively. By reducing the number of convolution kernels in the first layer of T-Net (Transformer Networks) to 128, the detection accuracy of cars and cyclists was increased by about 1% and 2% respectively, which confirmed that the model could effectively improve object detection accuracy.

    Table and Figures | Reference | Related Articles | Metrics
    Fast 4-points congruent sets for coarse registration of 3D point cloud
    Shiguang LIU,Hairong WANG,Jin LIU
    Journal of Shandong University(Engineering Science)    2019, 49 (2): 1-7.   DOI: 10.6040/j.issn.1672-3961.0.2018.244
    Abstract2242)   HTML146)    PDF(pc) (6934KB)(711)       Save

    In order to solve the problem that the 4-points congruent sets (4 PCS) method suffered from low computational efficiency and high registration errors when the overlap rate of two pieces of input point clouds was low, fast 4-points congruent sets (F-4PCS) was put forward. A new method for selecting four-point basis was presented. The source point cloud and target point cloud were given, their boundaries were separately extracted and extended as the boundary feature bands, and then a consistent four-point basis set was chosen from the boundary feature bands. This method could avoid some unnecessary iterations. By limiting the characteristics of the four-point basis, the invalid four-point basis was removed, it could reduce the verification time of the algorithm and improve the computational efficiency. Experiments results carried out on the relevant data sets showed that the F-4PCS method was more efficient than conventional 4PCS method in the case of low overlap rate of input point clouds and the registration success rate was higher than state-of-the-arts.

    Table and Figures | Reference | Related Articles | Metrics
    Chinese short text classification method based on word2vec embedding
    Mingxia GAO,Jingwei LI
    Journal of Shandong University(Engineering Science)    2019, 49 (2): 34-41.   DOI: 10.6040/j.issn.1672-3961.0.2018.197
    Abstract2119)   HTML49)    PDF(pc) (2868KB)(307)       Save

    In the short text classification process, the weak feature expression of the limitation of the number of words restricted the classification effect. To solve this problem, a Chinese short text classification method based on embedding trained by word2vec from Wikipedia (CSTC-EWW) was proposed, and a series of experiments for short texts with 4 topics from the website were finished. This method firstly trained the embedding by word2vec from Wikipedia corpus. the feature of short text based on the embedding was established. Naive Bayes and SVM was used to classify short text. The experimental results showed the following conclusions: CSTC-EWW could effectively classify short texts and the best F-value could reach 81.8%; Comparing the text feature expression of BOW model weighted by TF-IDF and the method of extending feature from Wikipedia, the classification results of CSTC-EWW were significantly better and F-measure of CSTC-EWW on car could be increased by 45.2%.

    Table and Figures | Reference | Related Articles | Metrics
    The improvement of the cochlear implantation CIS speech processing method based on the all phase DFT filter
    PANG Zhi-min,TIAN Lan,HOU Zheng-xin
    Abstract1969)      PDF(pc) (715KB)(2250)       Save
    Cochlear implantation is a new technology to restore the hearing ability for deaf people. A new type of digital filter with better filtrating ability, that is an all phase DFT filter, was introduced to simulate cochlear implantation CIS speech processing and to synthesize the voice, using the simple model of electrode excitation cochlear implant. By doing this, the voice can be simulated for the person with a cochlear implant. Comparisons between this model and the traditional Butterworth filter were made. The results of the experiment show that the synthesized speech spectra by the all phase DFT digital filter is closer to the original speech spectra, and is much better than the results of the Butterworth filter.
    Related Articles | Metrics
    Segmentation of connected characters based on improved drop-fall algorithm
    Qiyue SONG,Xuewen MU,Huan CHENG
    Journal of Shandong University(Engineering Science)    2018, 48 (6): 89-94, 108.   DOI: 10.6040/j.issn.1672-3961.0.2018.199
    Abstract1858)   HTML77)    PDF(pc) (3250KB)(649)       Save

    As the traditional segmentation methods could not segment connected characters correctly, a segmentation algorithm based on improved drop-fall algorithm was proposed. The algorithm included two steps. Zhang-Sueng's thinning algorithm and the clustering of the connected region via self-organizing maps was used to find the starting drop point of drop-fall algorithm. A new drop path was defined to improve drop-fall algorithm. The water dropped from the starting drop point, along the skeleton of the character overlap stroke, at the end of the overlapped stroke skeleton, then continued dropping along the slant angle direction of the skeleton, until met the boundary of the character connected part. The water drop path was defined as the connected character segmentation path. This method solved the problem of character strokes fracture caused by the traditional drop-fall algorithm. Compared with the traditional drop-fall algorithm and the vertical projection segmentation algorithm, the experimental results showed that it was an ideal method for segmenting connected characters.

    Table and Figures | Reference | Related Articles | Metrics
    Review of capsule network
    Jucheng YANG,Shujie HAN,Lei MAO,Xiangzi DAI,Yarui CHEN
    Journal of Shandong University(Engineering Science)    2019, 49 (6): 1-10.   DOI: 10.6040/j.issn.1672-3961.0.2019.312
    Abstract1703)   HTML105)    PDF(pc) (4331KB)(999)       Save

    Recently capsule network with dynamic routing was the new neural network model which was considered a significant model in next generation. In recent years, much research evidenced capsule network exceptional ability to fit features. But the high computational overhead made it unable to fit complex and large datasets. Consequently, reducing computational became a research hotspot. There were two methods, including optimized capsule and optimized routing, to solve the issue. Optimized capsule was usually driven by application purpose which was designed as a model of specific classification tasks. And optimized routing was the way to improve the performance of the model from an algorithmic perspective.

    Table and Figures | Reference | Related Articles | Metrics
    Fast scene recognition based on LDB descriptor and local spatial structure matching
    Dongbo ZHANG,Tao KOU,Haixia XU
    Journal of Shandong University(Engineering Science)    2018, 48 (5): 16-23.   DOI: 10.6040/j.issn.1672-3961.0.2017.409
    Abstract1617)   HTML33)    PDF(pc) (11270KB)(305)       Save

    A new local difference binary (LDB) descriptor and local spatial structure matching method was proposed to implement fast scene recognition. The multi-grid dense sampling method was used to obtain grayscale and gradient information of the image area, and the binary description was performed by comparing the grayscale and gradient size between the grids, which inherited the advantages of fast and low storage of binary feature extraction. The multi-point matching was adopted to replace the original single point of matching technology, which removed a large number of mismatches, thus the match accuracy was improved. The experiment showed that the efficiency of this method was about 2.7 times of SIFT and 1.9 times of SURF. The validity and recognition performance of the method were fully verified.

    Table and Figures | Reference | Related Articles | Metrics
    Comparison and analysis on measure indexes for structural hole nodes in social network
    HAN Zhongming, WU Yang, TAN Xusheng, LIU Wen, YANG Weijie
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2015, 45 (1): 1-8.   DOI: 10.6040/j.issn.1672-3961.1.2014.120
    Abstract1613)      PDF(pc) (2601KB)(2214)       Save
    In order to analyze different factors that affected structural holes measurement in social networks, seven existing methods to measure structural hole nodes were analyzed. Four groups of 12 simulated networks were built. Measure indexes for structural hole nodes were deeply and overall analyzed and compared in the simulated network, which were testified and analyzed in social network of Renren websites. The experimental results showed that seven existing methods perform poorly on identifying the structural hole nodes and some methods were highly correlated. Among these seven methods, betweenness centrality was relatively more effective.
    Reference | Related Articles | Metrics
    Early diagnosis and life prognosis for slowlyvarying fault based on deep learning
    ZHOU Funa, GAO Yulin, WANG Jiayu, WEN Chenglin
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2017, 47 (5): 30-37.   DOI: 10.6040/j.issn.1672-3961.0.2017.193
    Abstract1606)      PDF(pc) (6482KB)(989)       Save
    In order to overcome the shortcoming of traditional early fault diagnosis methods, a method of combining deep learning with PCA to realize early diagnosis of slowly varying small fault and life prognosis was proposed. Using the deep learning method to extract the sampled data characteristics layer by layer, learning the early fault characteristics and establishing the early fault diagnosis model for slowly varying small fault, the combining deep learning was combined with PCA to integrate the high dimensional fault feature vector extracted by the deep learning into a fault characteristic variable. A data-driven fault precursor could be defined according to the evolution rule of the characteristic variable of the historical fault data, and life prognosis model was established by exponential nonlinear fitting method. The TE benchmark data was used to verify the effectiveness of the proposed algorithm, experimental results showed the validity of the proposed algorithm by comparing with other algorithms.
    Reference | Related Articles | Metrics
    The time series prediction model based on integrated deep learning
    HE Zhengyi, ZENG Xianhua, QU Shengwei, WU Zhilong
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2016, 46 (6): 40-47.   DOI: 10.6040/j.issn.1672-3961.1.2016.213
    Abstract1567)   HTML    PDF(pc) (2496KB)(1767)       Save
    The conditional restricted Boltzmann machine time series model based on the Gaussian process(GCRBM)could efficiently predict single type of time series data, but the model could not make accurate predictions for multi-category data and real high-dimensional data. To solve the problem above, the time series prediction model based on integrated deep learning was proposed. Multiple deep belief networks(DBN)corresponding to the multi-category timing data was trained to study low dimensional feature. The low dimensional feature of multi-category data was used to train multiple GCRBM models. When the time series was predicted, the dimensionality of the model was reduced and categories of target data were identified by DBN model's reconstruction error, and the sequence of target data was predicted by the GCRBM model. The experimental results based on CASIA-A gait data set showed that the method could accurately recognize the categories of gait sequences and the predicting result could simulate the true gait sequences, which demonstrated the validity of the model.
    Reference | Related Articles | Metrics
    Images auto-encoding algorithm based on deep convolution neural network
    Yijiang HE,Junping DU,Feifei KOU,Meiyu LIANG,Wei WANG,Ang LUO
    Journal of Shandong University(Engineering Science)    2019, 49 (2): 61-66.   DOI: 10.6040/j.issn.1672-3961.0.2017.432
    Abstract1555)   HTML35)    PDF(pc) (1646KB)(505)       Save

    At present, image coding research was focused on information lossless, but it did not reflect the social network image differentiation. A novel social network images auto-encoding algorithm based on deep convolution neural network was proposed. The algorithm obtained good performance on image auto-encoding, which combined the feature extraction ability of deep convolutional neural network and characteristics of images in social networks. It combined the characteristics of the social network image with the clustering algorithm to cluster social network image and got the distance information, next the deep convolutional neural network was used to learn the distance information of these images, then it extracted the fully connected layer in the deep convolution neural network as the image coding, repeated the above steps and got the image coding finally. The experimental results showed that the proposed algorithm performed better than other algorithms of image search, and was more adaptive in the social network image search than that of the other algorithms mentioned.

    Table and Figures | Reference | Related Articles | Metrics
    Bound gait controlling method of quadruped robot
    MENG Jian, LI Yibin, LI Bin
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2015, 45 (3): 28-34.   DOI: 10.6040/j.issn.1672-3961.0.2014.328
    Abstract1548)      PDF(pc) (3687KB)(997)       Save
    Aiming at the problem of running control of quadruped robot, a running control method based on bound gait was proposed. The bound gait of the quadruped robot was implemented by fast and small range swing motion of the legs. A finite state machine was used to separate one complete cycle of motion into six stages, three stages for fore legs and three for hind legs respectively. In contact and buffering stage, vertical spring-damper model was used; in thrust stage, virtual model was used to adjust the thrust direction of the legs; and in swing stage, Bezier curve was used to plan the trajectory of the toes. By constructing a virtual model with the same size and mass with the hydraulic driven quadruped robot SCalf-II in the dynamics simulation software, the control method was verified and tested, simulation results showed that the robot came into a cyclic bounding motion with strong periodicity after five periods, the speed vibration in forward direction was small, the joint range of motion, speed and torque were all within the range of the design objective of SCalf-II, which verified the correctness and effectiveness of the proposed method.
    Reference | Related Articles | Metrics
    A microblog rumor events detection method based on C-GRU
    Lizhao LI,Guoyong CAI,Jiao PAN
    Journal of Shandong University(Engineering Science)    2019, 49 (2): 102-106, 115.   DOI: 10.6040/j.issn.1672-3961.0.2018.189
    Abstract1521)   HTML39)    PDF(pc) (1249KB)(505)       Save

    A microblog rumor events detection model based on convolution-gated recurrent unit(C-GRU) was proposed. Combining the advantages of CNN and GRU, the microblog event′s posts was vectorized. By learning the features representation of the microblog windows through the convolution layer of CNN, the features of microblog windows was spliced into a sequence of window feature according to the time order, and the sequence of window feature was put into the GRU to learn feature representation of sequence for rumor events detection. Experimental results from real data sets showed that this model had better ability to rumor detection than other models based on traditional machine learning, CNN or RNN.

    Table and Figures | Reference | Related Articles | Metrics
    A method on building Chinese sentiment lexicon for text sentiment analysis
    ZHOU Yong-mei1, YANG Jia-neng2, YANG Ai-min1
    Abstract1487)      PDF(pc) (1259KB)(5345)       Save

    A method on building Chinese sentiment lexicon based on HowNet and SentiWordNet was proposed,in which sentiment intensity of the word was automatically calculated by decomposing it into multiple semantic units and a lexicon proofreading technique was used to optimize the value of sentiment intensity of the word. The building lexicon was applied to the task of sentiment analysis, in which the support vector machine was used to build the sentiment classifier. The experiment results showed that the built sentiment lexicon was more effective than the general polar sentiment lexicon,and provided an effective dictionary resource for the research of sentiment analysis.

    Related Articles | Metrics
    Intelligent control of the discharge gap during UVEDM
    HU Yu-jing,ZHANG Jian-hua,REN Sheng-feng,BAI Wen-feng
    Abstract1476)            Save
    Ultrasonic Vibration Assisted Electric Discharge Machining is a 3C (Complex Plant, Complex Task, Complex Environment) system. The performance of control system is crucial for compound machining. A new FNN controller is developed and the learning algorithm is designed in this paper. The controller can learn it self by using the artificial neural network; the learning result can adjust the controlling rules with the condition changing, which optimize precision and improve the real time characteristic of the system. The correlation model between the machining performance and discharge parameters is established according to the experiment data, which is then applied to control the discharge gap. The experiment results indicate that the machining effect of the new FNN controller is better than that of the traditional machining tool.
    Related Articles | Metrics
    Milling force prediction model for highspeed end milling 3Cr2Mo steel
    WANG Su-yu,<\sup>,AI Xing<\sup>,ZHAO Jun<\sup>,LI Zuo-li<\sup>,LIU Zeng-wen<\sup>
    Abstract1467)            Save
    Milling force prediction model for highspeed end milling 3Cr2Mo steel
    Related Articles | Metrics
    Indoor wireless positioning based on ensemble deep neural network
    Dongdong SHEN,Fengyu ZHOU,Mengyuan LI,Shuqian WANG,Renhe GUO
    Journal of Shandong University(Engineering Science)    2018, 48 (5): 95-102.   DOI: 10.6040/j.issn.1672-3961.0.2018.169
    Abstract1411)   HTML27)    PDF(pc) (6449KB)(280)       Save

    Because of the low fault tolerance and weak anti-noise ability of fingerprint database in traditional wireless positioning model, an ensemble deep neural network wireless positioning method based on data fusion was proposed. This method could effectively overcome the interference caused by abnormal samples and noisy data on the wireless positioning system by sampling from the original fingerprint database randomly to generate train data for each base learner. During the process of fingerprint database construction, the Gauss-Occupied (G-O) data expansion method was proposed to solve the limitation of the small sample size of the wireless fingerprint database and decrease the cost of manual acquisition sharply, which increased the scope of the sample′s characterization. The results of the experiment showed that the proposed ensemble deep neural network wireless positioning model could not only improve the average positioning accuracy and the anti-noise ability of the wireless positioning system, but also reduce the maximum single point error in the positioning process.

    Table and Figures | Reference | Related Articles | Metrics
    Active driving behavior analysis algorithm based on monocular camera
    Chenmou WU,Zhijun FANG,Jenqneng HWANG
    Journal of Shandong University(Engineering Science)    2018, 48 (5): 69-76.   DOI: 10.6040/j.issn.1672-3961.0.2017.424
    Abstract1405)   HTML15)    PDF(pc) (9086KB)(453)       Save

    In order to prevent accidents, an algorithm for recognizing and monitoring the driver′s behavior based on the three-dimensional pose estimation of the human body was proposed. A monocular camera was used to capture the video stream of the driver in motion, the two-dimensional contour features of each frame of the image was extracted, and the two-dimensional projection of the pre-established three-dimensional human body model was matched to estimate the attitude of the driver′s upper body in real time. Based on the three-dimensional coordinates of the driver′s eight skeletal nodes, the driver′s behavior was identified and analyzed. Four driving states of driver′s normal, one-handed, answering calls and fatigue/drunk driving were simulated. Through the coordinate changes of the skeletal nodes, the gesture behavior of the driver could be detected and recognized, and the driver could be given reminders. When the light was enough, the algorithm could reduce the false detection rate by 24.24% compared with the PRECLOSE algorithm.

    Table and Figures | Reference | Related Articles | Metrics
    A method of DOA estimation by a special array structure
    CHEN Lei,WANG Jian-ying,LV Xue,WANG Feng
    Abstract1358)      PDF(pc) (253KB)(3015)       Save
    A new method of DOA estimation based on a special array geometry arrangement is presented. By adding a sensor at a specific location on uniform linear array, a group of special sensor pairs can be obtained, which can break the ULA steering vector's periodicity. The special array not only overcomes the weakness of the ambiguity of DOA estimation, but also achieves the higher resolution than ULA at the same hardware cost. A matrixblock spatial smoothed algorithm applied to the special array is proposed. The data covariance matrix is divided into some blocks according to its steering vector's periodicity, then the traditional spatial smoothed algorithm is applied to each block. After reorganizing the data covariance matrix is amended, the DOA estimation of coherent source can be achieved.
    Related Articles | Metrics
    Design of fuzzy control for traffic signals of a urban arterial intersection
    ZHAO Jian-yu,JIA Lei,ZHU Wen-xing,YANG Li-cai
    Abstract1329)            Save
    Urban arterial roads play an important role in urban traffic, and the control efficiency of the arterial roads intersection will affect the whole traffic of the city. A practical approach to fuzzy control algorithm of urban traffic flow of intersection is described. Based on existing fuzzy logic controller, a green light termination algorithm is added, which can adjust green light extension time to avoid unnecessary extension time. So that the split is optimized by fuzzy logic. A simulation experiment was performed to measure the capability of above fuzzy controller at a four-phase intersection, and the performance was measured with average vehicle delays through the intersection. The simulations show that the results of the fuzzy controller are more satisfactory than fixed time method.
    Related Articles | Metrics
    Lung tumor images recognition based on PSO-ConvK convolutional neural network
    Mengmeng LIANG,Tao ZHOU,Yong XIA,Feifei ZHANG,Jian YANG
    Journal of Shandong University(Engineering Science)    2018, 48 (5): 77-84.   DOI: 10.6040/j.issn.1672-3961.0.2018.191
    Abstract1329)   HTML11)    PDF(pc) (3246KB)(658)       Save

    In order to solve problems that convolution kernel was random initialization and the gradient descent method to train convolution neural network was easy to fall into local minimum, an image recognition method based on particle swarm optimization for convolution kernel was proposed. CNN(convolution neural network) was constructed by using the parameter migration method, and convolution kernel was extracted. The particle swarm algorithm was used to update the velocity and position of particles constantly and find the global optimal value to initialize convolution kernels. Convolution kernels were transferred to convolution neural network, and lung tumor images were used to train them. CNN model was trained by lung tumor images, and gradient descent method was used to modify network weights, hence global optimization ability of PSO algorithm was combined with local search ability of gradient descent method. The experiments verified effectiveness of method through three perspectives: batch sizes, iteration numbers, and network layer numbers. The particle swarm algorithm was compared with gauss function. The recognition rates of PSO optimized convolution kernel were always higher than that of randomized convolution kernel and gauss convolution kernel, its recognition rate reached 98.3%, which had certain feasibility and superiority.

    Table and Figures | Reference | Related Articles | Metrics
    Research of traffic signal optimized control algorithm based on fuzzy logic
    ZANG Li-lin,JIA Lei,LIN Zhong-qin
    Abstract1315)            Save
    Presents signal changed phase control algorithm for single intersection based on fuzzy logic and finishes the simulation. The traditional signal control algorithms are concerned about optimization of phases' time and ignore optimization of phases' combination. Take example for four-phase signal control used commonly in real world, research discovers that flexible combination of phases can acquire better control results under the premise of unchanged old order of four phases. Compared with traditional signal fuzzy control method for single intersection, signal changed phase control algorithm possesses flexible phase combination and phase time, it adapts to the real-time change of traffic status in the intersection very well. Simulation is finished based on the typical the four corners and traffic flow in different stages, the result indicates that signal changed phase fuzzy control is prior to traditional fuzzy control for the intersection of uneven traffic flow and it can reduce vehicle delay time greatly. The proposed algorithm is practicable and useful for city intersection signal control.
    Related Articles | Metrics
    Self-adaption acceleration slip regulation control of four-wheel independently-driving electric vehicle
    ZHANG Bohan, CHEN Zheming, FU Jianghua, CHEN Bao
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2018, 48 (1): 96-103.   DOI: 10.6040/j.issn.1672-3961.0.2017.081
    Abstract1305)   HTML    PDF(pc) (3899KB)(406)       Save
    There are many problems in the research of the acceleration slip regulation control, such as the single modeling method, the ideal slip rate of the controller can not match the changing road conditions, and the influence of the axle load transfer on the wheel adhesion performance is not taken into account. In order to solve the above problems, the vehicle dynamics model based on Carsim and Matlab/Simulink co-simulation was established. A self-adaption acceleration slip regulation controller based on double fuzzy algorithm was designed. The pavement recognition module could be added to the controller to estimate the changeable pavement attachment condition. Based on the estimation results, the optimal target slipping rate could be selected to realize the self-adaptive control. The acceleration slip regulation control algorithm of front-rear axle drive was designed separately to realize the differential control of front and rear axles. According to the different working conditions, the vehicle model and the acceleration slip regulation controller were verified.
    Reference | Related Articles | Metrics
    Suggestion sentence classification model based on feature fusion and ensemble learning
    Pu ZHANG,Chang LIU,Yong WANG
    Journal of Shandong University(Engineering Science)    2018, 48 (5): 47-54.   DOI: 10.6040/j.issn.1672-3961.0.2018.207
    Abstract1302)   HTML14)    PDF(pc) (1579KB)(341)       Save

    As an emerging research task, suggestion mining has gradually attracted attention of researchers in recent years. Compared with English language suggestion expression forms, those of Chinese were more abundant, and many different characteristics were present. It was necessary to carry out the research on suggestion mining in the Chinese environment. As suggestion sentence detection was the core task of suggestion mining, this research proposed an ensemble learning model that integrated the Stacking and Bagging methods to classify the reviews for the detection of suggestion sentence. The model firstly used Stacking to combine classifiers and constructed probabilistic feature space. Then, the convolution neural network (CNN) and paragraph vector (PV) model were used to construct the CNN feature space and paragraph vector feature space of the reviews respectively. Finally, the above features were fused and the Bagging classifier was trained to classify suggestion sentences. Experimental results on Chinese dataset verified the effectiveness of the model.

    Table and Figures | Reference | Related Articles | Metrics
    Cloud robotics: concept, architectures and key technologies
    TIAN Guohui, XU Yaxiong
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2014, 44 (6): 47-54.   DOI: 10.6040/j.issn.1672-3961.0.2014.282
    Abstract1297)      PDF(pc) (1959KB)(1761)       Save
    The technology of cloud robotics is the combination of cloud computing and robotics. It can bring great advantages in task execution and resource sharing for robots, and has become a hot topic in intelligent robot research field. In this paper, the developments of cloud robotics, key technologies, main platforms, main architectures and application prospects were comprehensively analyzed. Firstly, the concept arose of cloud robotics and its developments were introduced, and also the key technologies were given. Secondly, three main cloud service platforms were horizontally contrasted, and the main architectures of cloud robotics were analyzed. Finally, the application prospects for cloud robotics were comprehensively presented.
    Reference | Related Articles | Metrics
    A new burst assembly algorithm for the FRRenabled OBS networks
    ZHAO Yi,ZHAO Long,ZHANG You-zhi,QI Zhi-feng
    Abstract1296)            Save
    In order to meet the future need of realtime and quasirealtime applications, the OBS network design requires short latency time. FRR (forward resource reservation) algorithm is an efficient method to reduce the endtoend data burst delay, but it still has some defections in bandwidth availability. Aiming at the potential problems in the FRR algorithm, PAFRR (precision assembly with forward resource reservation) is proposed. Both theoretical analysis and simulation results demonstrate the advantages of the PAFRR over to the FRR scheme.
    Related Articles | Metrics
    A variable structure cascade double loop control to the startup process of the heavy truck
    LI Hong-bin,ZHANG Cheng-rui
    Abstract1294)            Save
    In order to enhance the smoothly character of the startup process of the heavy truck, a startup control algorithm is proposed based on the analysis of the acceleration fluctuation before and after the lock point, which adopts variable structure cascade double loop control to keep the rotating speed of the engine constant. Considering the nonlinear characteristics of the clutch operating system, the neuron adaptive PID control algorithm is adopted in the acceleration loop to solve the problem of nonlinear control. The experiment results show that acceleration fluctuation before and after the lock point can be restrained by using this algorithm, then the performance of the launch process is improved.
    Related Articles | Metrics
    A prediction method of atmospheric PM2.5 based on DBNs
    ZHENG Yi, ZHU Chengzhang
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2014, 44 (6): 19-25.   DOI: 10.6040/j.issn.1672-3961.1.2014.180
    Abstract1289)      PDF(pc) (1805KB)(2560)       Save
    A DBNs-based (deep belief networks) method for forecasting the daily average concentrations of PM2.5 in Xian was proposed. Besides, the way to select training data set as well as the DBNs parameters was optimized. Then relative experiments and comparison with methods based on BP (back propagation) and RBF (radial basis function) artificial neural network confirmed the feasibility and precision of DBNs. The results showed that the MSE (mean square error) between DBNs simulated PM2.5 daily average concentrations and observed ones was 8.47×10-4 mg2/m6, while the MSE of RBF and BP was 1.30×10-3 mg2/m6 and 1.96×10-3 mg2/m6 respectively. Therefore the DBNs-based method was fit for prediction of PM2.5 concentrations and it predicted more accurately than those methods based on RBF and BP artificial neural network.
    Reference | Related Articles | Metrics
    An automatic reading method for pointer meter
    Yanghao ZHOU,Yifan LIU,Li LI
    Journal of Shandong University(Engineering Science)    2019, 49 (4): 1-7.   DOI: 10.6040/j.issn.1672-3961.0.2018.275
    Abstract1288)   HTML151)    PDF(pc) (4120KB)(683)       Save

    An automatic reading method for automatically monitoring pointer meter in substation was proposed based on the machine learning and image processing algorithms, which was consisted of two stages: meter detection and pointer recognition. The position of the meter in the input image was detected by using the fully convolutional networks, and then the patch of the meter was extracted. The interference of illumination and shadow on the pointer recognition was reduced by using histogram equalization, median filtering and bilateral filtering, and the tilt of shooting was rectified by using the affine transformation. The position of the pointer was detected via the improved Hough transform. The reading was obtained by computing the angle of the pointer. The results showed that the method could detect the pointer meter and recognize the reading accurately for the pointer instrument in the substation. The method showed good robustness to the disturbances such as illumination and shadow, which could significantly reduce the substation inspection personnel workload and improve the work efficiency.

    Table and Figures | Reference | Related Articles | Metrics
    Chinese entity relation extraction based on entity disambiguation
    SHAO Fa, HUANG Yinge, ZHOU Lanjiang, GUO Jianyi, YU Zhengtao, ZHANG Jinpeng
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2014, 44 (6): 32-37.   DOI: 10.6040/j.issn.1672-3961.1.2014.163
    Abstract1277)      PDF(pc) (1979KB)(1544)       Save
    To solve the polysemy problem in Chinese Entity Relation Extraction in open text, a Chinese entity relation extraction method based on entity disambiguation was proposed. First, mining entity relation pairs from HowNet,and the entities were mapped from HowNet to Wikipedia by using disambiguation method based on Bayesian classification so as to obtain high-quality relationship instance; Then, extracting the sentence instances in the corresponding context with these relation instances, to construct a basic extraction pattern; Finally, extracting new cases use the new pattern. The experimental results showed that the accuracy of the proposed method was higher than the methods without semantic disambiguation and pattern merging.
    Reference | Related Articles | Metrics
    Construction of index system for relay protection operation evaluation based on KPI
    ZHENG Maoran, YU Jiang, CHEN Hongshan, ZHANG Shiyu, CHENG Haoyuan, GAO Honghui, ZHANG Jingwei
    JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE)    2017, 47 (6): 13-19.   DOI: 10.6040/j.issn.1672-3961.0.2017.225
    Abstract1276)      PDF(pc) (710KB)(312)       Save
    A method to construct the evaluation index system of relay protection operation based on KPI was proposed and the index system from three layers was illustrated. According to 20-80 rules and the balanced score card method in KPI theory, the composition of the strategic indicators was clarified and the strategic index set was established after analyzing the relationship between the relay protection business and other related business of power grid. The strategic index decomposition method based on the fishbone map was established and tactical sub-indicators was formed through the dimension division, classification design and index integration constructing tactical indicators set. Strategic indicators set and tactical indicators set were closely integrated in the form of index chain. The relay protection operation evaluation index system consist of multi-layer, multi-level and multi-dimension was finished.
    Reference | Related Articles | Metrics