Electric, Electronic and Control Engineering 9781138028425_FM.indd Tai ngay!!! Ban co the xoa dong chu nay!!! 6/9/15 7:16 PM PROCEEDINGS of the 2015 International Conference on Electric, Electronic and Control Engineering (ICEECE 2015), Phuket Island, Thailand, 5–6 March 2015 Electric, Electronic and Control Engineering Editors Fun Shao Digital Library Department, Library of Huazhong University of Science and Technology, China Wise Shu Art College, Hubei Open University, China Tracy Tian Bos’n Academic Service Centre, China 9781138028425_FM.indd 6/9/15 7:16 PM CRC Press/Balkema is an imprint of the Taylor & Francis Group, an informa business © 2015 Taylor & Francis Group, London, UK Typeset by diacriTech, Chennai, India Printed and bound in China by CTPS DIGIPRINTS All rights reserved No part of this publication or the information contained herein may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, by photocopying, recording or otherwise, without written prior permission from the publisher Although all care is taken to ensure integrity and the quality of this publication and the information herein, no responsibility is assumed by the publishers nor the author for any damage to the property or persons as a result of operation or use of this publication and/or the information contained herein Published by: CRC Press/Balkema P.O Box 11320, 2301 EH Leiden, The Netherlands e-mail: Pub.NL@taylorandfrancis.com www.crcpress.com – www.taylorandfrancis.com ISBN: 978-1-138-02842-5 (Hardback) ISBN: 978-1-315-67504-6 (eBook PDF) 9781138028425_FM.indd 13/06/15 10:23 AM Electric, Electronic and Control Engineering – Shao, Shu & Tian (Eds) © 2015 Taylor & Francis Group, London, ISBN 978-1-138-02842-5 Table of contents Preface xiii Acknowledgement xv Organizing committee xvii Adaptive algorithm of parallel genetic optimization based on orthogonal wavelet of space diversity X.G Pei & S.H Zheng Research and application of comprehensive evaluation and detection analysis platform of transformer bushing state A.Q Cai, Q.L Zhang, Y.W Li & Q Zeng Designing a state transition circuit with Creator2.0 and its PSoC realization Y.P Liu, W Wang, J.D Huang, X He, T Huang & B.K Liu 11 An obstacle avoidance scheme for autonomous robot based on PCNN T Xu, S.M Jia, Z.Y Dong & X.Z Li 15 A study on thematic map of power grid operation based on GIS Y Liu, Z.H Cui, D.Y Wu & J.L Zhao 21 Capacity ratio relation of energy storage and intermittent DG and practical estimation method Y Zhang, F.Y Yang, J Zeng, G.Y Zou & L Dong 29 Design of English educational games and SPSS analysis on application effect Y.H Sun 35 Design and FEA of a light truck two-leveled leaf spring H.J Wu, X.F Dai & Z.X Zheng 43 A brief discussion on research and application of intelligent management and control alarm platform of unattended transformer F.C Li, K Hu, H Zhang, B Song & Z.J Hu 49 Research and implementation of yoga system based on virtual reality S.J Li & Z.D Xiong 53 On fusing substation video surveillance with visual analysis under integrated dispatch and control J.L Zhang, P.L Chen, L.J Feng, M.D Li, X.Q Zhao & Y Liu 61 A study on the system of teaching quality based on classification rule F Cui 67 Design of a high efficiency 2.45-GHz rectifier for low input power energy harvesting Q.Q Zhang, H.C Deng & H.Z Tan 73 Investment portfolio model design based on multi-objective fuzzy comprehensive evaluation method Z.X Shang, Y Wang & X Liu 77 Design and implementation of the system of impact location based on acoustics detecting technique D.H Fang & H.T Jia 81 Application in robot of the three-dimensional force tactile sensor research based on PVDF Q Pan, Z Wan & S.L Yi 87 v 9781138028425_FM.indd 6/9/15 7:16 PM Application of auto-control in car repair X.J Yuan & Z.N Lu 93 The development history of engineering cost consulting industry in mainland china K.C Li, J.H Yan & J Ding 99 Insulator ESDD prediction based on least-squares support vector machines H.L Li, X.S Wen & N.Q Shu 103 A Reliable Privacy-preserving Attribute-based Encryption T.F Wang, L.J Zhang & C Guo 109 Research on the agility of C2 organization S Chen, X.J Ren & W.J Shao 115 Experimental research and analysis on acoustic emission from polluted insulator discharge H.L Li, N.Q Shu & X.S Wen 119 A Fault location method for micro-grid based on distributed decision F Tang & W Gao 125 Application of computer automatic control technology in the industrial production site Z Lv 133 Risk evaluation for wind power project based on grey hierarchy method C.B Li, K Zhao & C.H Ma 139 Automatic control system of automotive light based on image recognition Z.N Lu & X.J Yuan 143 The research and application of in vitro diagnosis using smart grid dispatch data Y.X Liao, H.B Li, H.X Yu, F Li & Z.L Mu 149 Design and implementation of a novel parallel FFT algorithm based on SIMD-BF model S.C Zhang, Y.H Li, Y Li & B Tian 153 Research and implementation of cellsense biosensor based on environmental engineering Q.L Liu 157 New construction methods for the optimum grassmannian sequences in CDMA L.X Wang, H Hu & J.B Yang 163 Research and application of data mining technology and operational analysis of the integration of power grid regulation and control D.Y Wu, J.L Zhao, Z.H Cui & Y Liu A study on the city coordination scheme of industrial integration Y.G Qi Power system black-start recovery subsystems partition based on improved CNM community detection algorithm Y.K Liu, T.Q Liu, Q Li & X.T Hu 169 173 183 Discussion on the application design of vertical greening in urban public spaces D.J Long & D Wang 191 New designation to multi-parameter measurement system based on chemical oscillation reaction X.N Chen, H.Y He, T Zhang, H.T Dong & D.X Zhang 197 Power electronic circuits fault diagnosis method based on neural network ensemble T.C Shi & Y.G He 201 The model, simulation and verification of wireless power transfer via coupled magnetic resonances H.G Zhang & M.X Zhang 207 Spatial distribution characterization of A-grade tourist attractions in Guizhou province by GIS H.L Fu 213 vi 9781138028425_FM.indd 6/9/15 7:16 PM A novel method for fundamental component detection based on wavelet transform H Chen & Y.G He 219 Low-noise, low-power geomagnetic field recorder H.F Wang, M Deng, K Chen & H Chen 225 Development of calibration module for OBEM H Chen, H.M Duan, M Deng & K Chen 229 Novel sampling frequency synchronization approach for plc system in low-impedance channel Y Wang, Z.L Deng & Y.Y Chen 237 Research and application of distribution network data topological model J.J Song, J.L Guo & Y.H Ku 241 Micro-grid storage configuration based on wind PV hydro-storage comprehensive optimization B Cao & Y Yang 245 Analysis model and data-processing method on vertical spatial characteristics of ship-radiated noise Y.S Liu & X.M Yang 253 Research of measuring technology of pulse current measurement based on embedded systems Y.B Yang, X Chen, H.O Yan, J Sun, M.Z Wang, X.L Wang, X.Y Ai, X.F Huang, Y Gao & X.Y Yang 259 Application research of improved PSO algorithm in BLDCM control system P.F Yan, Y.L Hu, X.L Zheng, C Yin, H.G Zang, Y.W Tao & G.L Li 263 Passivity-based control for doubly-fed induction generator with variable speed and constant frequency in wind power system J.R Wang & P.P Peng 267 System structure of IEC61850-based standard digital hydropower station R.W Lu, Y.H Gui & Q Tan 273 Information hiding method based on line spectrum frequency of AMR-WB D Teng, H.N Feng & J.J Yu 279 CMOS-Integrated accelerometer sensor for passive RFID applications M.X Liu, Y.G He, F.M Deng, S Li & Y.Z Zhang 285 Ship course control based on humanoid intelligent control Y Zhao, Y.L Zhao & R.Q Wang 293 Integrated evaluation method for transmission grid safety and economy and its application H.J Fu, H.Y Wang, J Chen, G.C Xue & Z.L Li 297 Improved digital image steganography algorithm X.D Wan & R.E Yang 301 Characteristic of HV insulators’ leakage current on surface discharge Y Tian, L.J Feng, T.Z Wang & M Jiang 307 New assessment method of internal model control performance and its application L Liu & Q Jin 313 The research on how to control a three-phase four-leg inverter T Jia, W.G Luo, N Liu & Y Yang 319 Hybrid anti-collision algorithm based on RFID system J.A Zhang, Y.G He, H Chen & M.X Liu 323 Research on disaster recovery policy of dual-active data center based on cloud computing X Chen & L.J Zhang 329 Defect identification in pipes by chirp signals F Deng & H.L Chen 333 vii 9781138028425_FM.indd 6/9/15 7:16 PM A comprehensive evaluation of investment ability of power grid enterprises–Taking power grid enterprise of Zhejiang province as an example J Fan, D.X Niu, X.M Xu, H.H Qin & H Xu 339 Research and application of online grid load monitoring and smart analysis system H.G Wang 343 A Study on sports video analysis based on motion mining technique X.P Chi 353 Study on the influence of two-navigation holes cross-sea bridge construction on tide environment of the existing channel C.Y Wang & Z Liu Simulation analysis for heat balance of groundwater heat pump in multi-field coupling condition X.Y An, W.D Ji & Y Zhao Research on lighting withstand performance of hvdc power transmission line based on the new electrical geometry model X.G Gao, J.Q Du, K.X Liu, X.Y Xie & Y Yuan 359 365 373 Design and development of online photoelectric detection turbidimeter for water environment H Zhang, Y.W Huang, Y Yu & B Xu 379 Implementation of accurate attention on students in classroom teaching based on big data Y.W Zheng, W.H Zhao, H.X Chen & Y.H Bai 383 A neural network method for measuring plate based on machine vision Y Liu, R.J Yang, Y.H Wang & J.Y Li 389 Framework study of accounting query and computing system under heterogeneous distributed environment Y Jia 393 LABVIEW-based simulation training system of chinese medicine bone-setting manipulation H.Y Mo, J Liu, H Cao, C Ni, J.Z Zhang & X.R Song 399 Object detection based on gaussian mixture background modeling J Liu, D.W Qiu, H Cao, J.Z Zhang & H.Y Mo 405 Design and implement of monitoring system for marine environment based on zigbee 411 L.C Wang & Z.Y Liang RFID sensor network for rail transit intelligent supervisory control S.S Yu & Y.G He 417 An algorithm of digital image denoising based on threshold Y.W Shi & R.E Yang 421 A Study on the construction status of smart power distribution network planning and the method of improvement 425 M Qi, N Zheng & L Peng An approach for improving the resolution of MUSIC W Feng, Y.W Wang, Z.Z Li & Z.L Wang 429 Crashworthiness optimization design of triangular honeycombs under axial compression Q He & D.W Ma 433 Research on service matching of single resource in cloud manufacturing M Zhang, Y.L Shang, C.Q Li, Y.H Chang & N Zhang 439 Research on cloud storage technology of a grouting monitoring system based on the Internet of things S Gao, H Zhang, Y.W Huang & X.W Yu 443 Object tracking algorithm based on feature matching under complex scenes J Liu, D.W Qiu, H Cao, J.Z Zhang & H.Y Mo 449 viii 9781138028425_FM.indd 6/9/15 7:16 PM Wide area backup protection algorithm based on fault area detection S Li, Y.G He & M.X Liu 455 Fuzzy comprehensive evaluation of university network security risks T Li 461 Analysis on incremental transmission loss and voltage level of wind power system with doubly-fed induction generators (DFIGs) T Wang, Y.G He & Mingyi Li 467 Research and prevention of illegal intrusion of digital television network X.F Hu 473 A study on numerical control transformation of milling machine based on interpolation algorithm C Sha & J Luo 479 Study on HCI of mobile internet—take graphic design of IOS7 windows as an example F Chang 487 Research on data acquisition for medical air quality detection Z.G Liu & B.Q Wang 491 A new type of photonic crystal fiber with high nonlinearity and high birefringence based on microstructure fiber core Q.C Meng & Y.K Bai 495 The application of LABVIEW-based virtual instrument technology in electro-hydraulic servo test system G Zhao, J.J Shi, L.H Sun & X.D Wang 501 The design and implementation of virtual oscilloscope based on LabVIEW L.H Sun, H Bai, G Zhao & J.F Ma 507 High level semantic image retrieval based on Ontology S.Q Wang, G.W Xu, C Zhang, B.B Chen & C.X Xu 515 A new method of pulse comparison detect in IR-UWB ranging L Zhu, H Zheng & H Zhang 519 Research on the influence of wind power integration on transmission service price H.J Kan, L.N Tan & Y.J Zhang 523 A recoverable color image blind watermark scheme and its application system based on internet X.L Chen & G.Q Hu 527 A new high-frequency transmission line ice-melting technique X.G Gao, Y.T Peng, K.X Liu, X.Y Xie & C.Y Li 531 A consistency evaluation and maintenance method of electric vehicle Lithium-ion+ battery based on resistance Y Xu & Y Yang 537 A machine learning practice to improve the profit for a Chinese restaurant X.B Li & N Lavrac 543 The control system research of X-ray generator in medical diagnostic Y Liu, B.B Dong, J.J Yang & B.Z Guo 551 Study on the magnitude-frequency response for RLC series circuit H.Y Zhou, L.R Li & Y.H Xie 557 Research on grey neural network based on genetic algorithm used in the air pollution index model B.B Chen & H.R Wu 561 Study on the influence of the cooperation network based on the PageRank algorithm B.B Chen, H.R Wu & G.W Xu 567 ix 9781138028425_FM.indd 6/9/15 8:30 PM Palmprint recognition using SURF features R.S Geng, X.J Tao & L Lei 573 A BTT missile optimal controller design based on diffeomorphism exact linearization C.Z Wei, C Cheng & Y.B Gu 579 A low power consumption indoor locating method research based on UWB technology H.P Hong, C Shen, Y.H Zhu, L.Q Zheng & F Dong 587 Research on unmanned ground vehicle following system Z.H Wang, N.Y Li, Y Zhang & J Zhang 591 The display and application of campus energy consumption of temperature and humidity data acquisition system X.L Wang, Y Yu & C Jiang Design and implementation of vulnerability database maintenance system based on topic web crawler H.Y Liu, Y.F Huo, T.M Xue & L.Q Deng A new method of high-accuracy detection for modal parameter identification of al parameter identification of power system low frequency oscillation Y Zhao, Z.M Li & T.Y Li 597 603 609 Bifurcational and chaotic analysis of the virus-induced innate immune system J.Y Tan, G.L Qin & Y.L Xu 615 A Multiple Utility Factors-Based Parallel Packet Scheduling Algorithm of BWM System M Wang, Q.Y Sun, S.G Zhang, Y Zhang & Y.L Liu 623 A new model for automatically locating the perceptional cues of consonants F Bai 629 A Detection method based on image segmentation applied to insulators Y.J Zhai, Y Wu & D Wang 633 Service restoration strategy of active distribution network based on multi-agent system H Ji & L.W Ma 637 A research and implementation of the automatic synchronization strategy under difference frequency power grid based on fuzzy control principle H Hu, J.Y Zhou,Y.G Fu, L.L Zeng, Y.L Li, Y.P Yang & Y.F Yue A method of coal identification based on D-S evidence theory T Wang, L Tian & W.N Wang Research on Green Spline interpolation algorithm application in optical path computation in aerodynamic flow field Y Zheng, H Sun & Y Zhao 645 651 657 A digital image steganography with low modification rate X.D Wan & R.E Yang 663 A wireless method for detection of half-wave direct current H.F Zhang, Z.G Tian & E.P Zhang 667 Design and implementation of intelligent toolbox based on RFID technology M.L Wang, Y.N Li, Y.P Teng & J Tan 673 Design and implementation of the control system in the underwater hydraulic grab H Zhang 679 Analysis on the pier influenced by sea dike causeway bridge T Yue, D.J Zuo, Q.S Mao & J Zhang 685 Study on influence of new-built sea dike on submarine optical fibre cable Y.J Luo & D.Y Tang 691 x 9781138028425_FM.indd 10 6/9/15 7:16 PM of the information contact mode of s ubsystems, the structural mode and the overall structure Therefore, improvements can be realized by the combination in such aspects as the effect of advantage complementation among algorithms For i=l: n For j=l: n %for each edge End Eod End 2.4 The improvement model of the swarm intelligence optimization The possibility of ants choosing k cities is that the provided heuristic information and the residual information content on the path from the city of ant to the target city are in the relation of a certain function It is a complete cyclic process after all the ants having passing through all n cities in order to prevent excessive submerging of information So, the information concentration on the path ij at (t+1) is: m τ ij (t + 1) = (1 − p).τ ij (t) + ∑ ∆Tij (t) (4) k =1 Here, the path of k ants at the moment t is represented byTk (t), the length of which isLk (t) Q is the related adjustment parameter and Q>0 The optimization path of TSP can be thusly obtained, corresponding codes of which are shown below [5]: %Ant system for solving the traveling salesman problem; m=n %Number of ants is equal to the number of towns For i=1: n % for each edge For j=l: n If i=j η(i.j)=1/D (i.j) % Visibility τ (i,j)= τ o % pheromone Elaeτ (i,j)=0 End lf End For End For For k=l: m End ; %Main; loop For t=1: tmax % tmax-number of iterations Fork=1: m % for each ant: End If”Is the best solution found? =”End 3 Applications of the swarm intelligence optimization algorithm in chemistry and chemical engineering Applications of optimization problems are widespread in the chemical engineering field An optimization algorithm with excellent performance is always the pursuit in the chemical engineering field The swarm intelligence optimization algorithm is highly efficient globally, the seeking ability of which is stronger than that of other algorithms The swarm intelligence algorithm has attracted much attention due to this feature Applications of the swarm intelligence optimization algorithm in chemistry and chemical engineering are introduced simply as follows 3.1 The application of the swarm intelligence optimization method in the exploitation of catalyst The method of logical operation is the core content of the computer-aided catalyst design But this method has such deficiencies as low convergent speed and low optimization efficiency that impede the rapid development of computer-aided catalyst design The swarm intelligence optimization method has certain advantages in complex optimization problems so that studies on catalyst design are hot topics There are still lots of work needed to be done 3.2 The application of the swarm intelligence optimization method in chemical metrology P S Shelokar[6-7] compared the ant colony algorithm with the genetic algorithm and the taboo search in terms of results The ant colony algorithm is able to forecast data collection and reveal the internal law, quality and processing time of which are relatively reasonable and practical Yaping Ding[8-9] et al proposed the chemical ant colony algorithm through the spectrum analysis with the swarm intelligence method It can be known from the comparison that the convergent speed is 40% faster than that of the genetic algorithm Besides, this method can be applied in systems of multiple groups like black, white and gray 740 9781138028425_Chp_244.indd 740 07/05/15 2:49 PM 4 Conclusion and prospect 3.3 The application of the swarm intelligence optimization method in the extraction process A I Papadopoulos et al composited heptanes with the genetic algorithm, the simulated annealing algorithm and the swarm intelligence optimization algorithm, and extracted the extraction agent composed by water, methylbenzene, dimethylbenzene, acetic acid, chloroform, acetone, etc CPU provides the operating time, which can be regarded as the evaluation standard It can be finally concluded that the advantage of the swarm intelligence optimization algorithm is relatively prominent However, the simulated annealing algorithm is very reliable in terms of robustness Standard deviations of all the objective functions are the smallest 3.4 The application of the swarm intelligence optimization method in chemical engineering analysis It was first put forward by V K Jayaraman that the swarm intelligence optimization algorithm can be applied to solving chemistry and chemical engineering problems They proposed a swarm intelligence algorithm in a paper, the global optimization design of which is an intermittent chemical engineering process with intermittent changes It is difficult to handle nonlinear mathematical models according to previous experience V K Jayaraman et al maintained the relatively stable diversity of ant colonies by taking advantage of variation characteristics and appropriate hybridization and solved the model of combined optimization problems, the production constraints of a single product, continuous function optimization model of intermittent production of multiple products, as well as combined optimization problems of intermittent production scheduling The optimization result with relatively fierce competitions is 0.573, which is extremely close to the iteration result 0.5735 The yield of intermediate products is 0.03% smaller than the optimization result of the least square method in terms of error Thus, Yijun He et al constructed a continuous ant colony optimization system (CACS) for constraint problems through the method of searching food so as to obtain the ant colony system of continuous optimization problems If the system model is applied in the preparation of butenoic alkane, the optimal solution of CACS is more optimized than that of α BB (α -based branch and bound ) Now, σ =5*10exp (-4) and the constraint violation is much smaller than that of α BB However, when this method is inferior toα BB, σ =5*10exp (-6) orσ =0 This process is not in violation of the constraints, but the final solution of α BB is in severe violation of the constraints 1 The swarm intelligence algorithm is a new bionic algorithm, establishing an approximate model in line with the behavior of ant foraging for spatial problems In the meantime, heuristic operators are introduced to improve the practicability The algorithm has a relatively low requirement for hardware, which is convenient for application and promotion 2 The swarm intelligence algorithm has the same effect with the genetic algorithm and the annealing algorithm in terms of applications in the chemical engineering field However, this algorithm features high robustness in chemistry and chemical engineering processes and provides new solutions for studies on chemical process optimization and chemical metrology 3 Because of the late starting, the swarm intelligence optimization algorithm is still in a primary stage compared to the simulated annealing algorithm and the genetic algorithm and still lacks strict mathematical proof It is a major research direction of applications of the swarm intelligence optimization method in chemistry and chemical engineering that most parameter can be acquired from experiments and experiences 4 The swarm intelligence optimization method is still in the exploratory phase in terms of application in chemistry and chemical engineering With the solution of a series of problems, the algorithm plays an important role not only in chemical engineering and chemical clustering but also in chemical difficulties References [1] Wei, P & Xiong, W.Q (2002) An ant colony algorithm for function optimization, Computer Science, 29(9):227–2291 [2] Cheng, B (2007) The Improvement of Two Random Optimization Algorithm and Applications in Chemical Engineering, Zhejiang University, 6.7:5–11 [3] He, Y.J (2008) The Swarm Intelligence Optimization Method and Its Application in Chemistry and Chemical Engineering, Zhejiang University, 1.1:4–10 [4] Yan, C.Y., Zhang, Y.P & Xiong, W.Q (2007) A new information content update strategy of the ant colony optimization algorithm, Application Research Of Computers, 24(7):86–88 [5] Du, L.F & Niu, Y.J (2011) The implementation of the ant colony algorithm, Information Technology, [6] Wu, Q.H, Zhang, J.H & Xu, X.H (1999) The ant colony algorithm with variation features, Computer Research and Development, 36(10):1240–1245 741 9781138028425_Chp_244.indd 741 07/05/15 2:49 PM [7] Shtovba SD (2005) Ant algorithms: theory and applications, Programming and Computer Software, 31(4):165–176 [8] Mitra K, DebK, GuPta S K (1998) Multi-objective dynamic optimization of an industrial nylon semibatch reactor using genetic algorithm, Journal of Applied Polymer Science, 69:67–85 [9] Ravi G GuPta S K, Ray M B (2002) Multi-objective Optimization of Cyclone Separators Using Genetic Algorithm, Industrial & Engineering Chemistry Research, 39:4270–4285 [10] Srinivas N, Deb K (1995) Multi-objective function optimization using non-dominated sorting genetic algorithms, Evolutionary Computation, 2(3):222–245 [11] Zitzler E, Thiele L (1999) Multi-objective evolutionary algorithms: a comparative ease study and the strength pareto approach, IEEE Transaction on Evolutionary Computation, 3(4):256–170 [12] Wang, Z., Liu, G.Q & Chen, E.H (2009) The K means algorithm for the optimization of the initial center, Pattern Recognition and Artificial Intelligence, 22(2):298–305 [13] Song, L., Li, M.Y & Li, X.Y (2008) The K means algorithm based on particle swarm optimization and its applications, Computer Engineering, 34(16): 201–203 [14] Wu, Y.W & Hu, X.G (2007) An optimization scheme of the K means algorithm, Journal of Chaohu University, 9(6): 21+24 742 9781138028425_Chp_244.indd 742 07/05/15 2:49 PM Electric, Electronic and Control Engineering – Shao, Shu & Tian (Eds) © 2015 Taylor & Francis Group, London, ISBN: 978-1-138-02842-5 The research of the long-distance runner’s physical index based on data analysis Cheng Liu Pingdingshan Institute of Education, Pingdingshan Henan, China Abstract: With the development of information technology, database technology has been widely used in the database management in the last two years The technology of data mining is the most advanced technology which has been widely used in the large-scale companies in the area of communication, bank, transportation, insurance, and so on The paper analyzes the physical indexes of the long-distance runners and the theory of data mining in order to solve the problems If we provide the right data to decision-makers, a better plan will be made in advance Keywords: data mining; athlete; arithmetic; physical index With the rapid development of science and technology, great changes have been taken in the field of sports training, and scientific training emerged as an important tool of the sports training, which had a positive impact on the athlete’s good performance Physical test of the athlete seems to be indispensable, so we can find much important information and regular pattern by analyzing the data of the test The coach could develop a scientific training program for long-distance runners according to this data Above all, data mining plays an important role in this process 1 Data mining technology 1.1 The meaning and task of data mining Data mining is a method of data analysis Finding out the useful data seems like extracting the iron from the ironstone, so we should analyze the large amount of data Data mining is a technology to extract the most important things from the original data, which seem to be valued, unknown and implicated The type of knowledge is called data mining tasks, and its main part is to summarize the rule of mining, classify the data and so on statistics, which mainly contain traditional statistical parts and systems of subjective guide The intelligence of data mining can be achieved based on the genetic algorithms and neural network techniques, and self-organizing networks and former spy network have been widely applied in the neural network In the field of engineering, decision tree is a simple method to express things which will be gradually separated into different areas and represent different categories The sub-branch will help to establish a decision tree 1.3 Classification and application of data mining The classification of data mining are not all the same, it is based on different type of mining, and it can be classified on the basis of knowledge such as the character rules and deviated rules It can also be classified by the usage, such as data driven mining, spontaneous knowledge mining, etc Based on the diversity, there exist two kinds of database: deductive databases and media databases The mining technology had been widely used in the market via data collection It will confirm the consumers’ habits in a particular way, which will predict the next step of them We will gain more money by analyzing the information on them 2 1.2 The methods and techniques of data mining Different data mining tasks require different methods and means On the whole, it can be divided into the following categories: statistical analysis, data mining, decision trees and rough sets Statistical analysis is to analysis the data by using the traditional 2.1 The process of data mining The process of the data mining refers to the way we get the information from the big data, which contains lots of aspects such as data collection, the assessment of the mode, etc In this paper, we take the mode of 743 9781138028425_Chp_245.indd 743 07/05/15 2:49 PM John as an example, which has brought some simple introductions of this process The model attaches more importance to personnel as well as experts of data mining during the whole process, so that experts can describe it clearly because of sufficient background knowledge, and the scholar will pay more attention to practical techniques and problems 2.2 Functional analysis of the TENNIS-DAMS TENNIS-DAMS system can solve the problems of the long-distance runners by analyzing the indexes The management issues consist of three parts, like management & analysis of data test It will analyze and classify the associated functions, and the data mining process can be roughly divided into stages such as data collection and management, data preprocessing, data mining (pattern discovery) and pattern processing and evaluation phase The original database has been shaped by the collection of data, and this stage is separated into input processing and maintenance module of data The stage of data preprocessing includes the comparative analysis and query module The goal of data mining in this system is to discover the relationship between the physical indicators and the physical state of the athletes, which may be the most important part in this article 2.3 Logical structure of system and the storage of database In the system, the database must be established, for which the database is the central function, and it is also the interface among the functional modules The module of the system has linked to each other by using the database, so that the respective functions could be completed (see Figure 1) In the process of mining, you can create the target database and schema libraries by yourself according to the reality, and it will be cost savings, easy management, query and analysis 3 Main rules and algorithms of data mining 3.1 The association rule mining kernel The concept of association rules is to make I = (i ,, i: , Iă , i} as a set of attributes of possible values, which is called the set of data items, and i- (1≤k≤n) is called data item to record a property’s value in database The number of elements in the database was called the length of the data item, while the length of the data item is called the n-dimensional Attribute association rules can be described by confidence, support, hope and confidence Suppose D contains C% and Y, C is considered as the confidence of the association rule X≥Y, as C%=The number of Transactions(X ∪ Y)/ The number of Transactions(X); Suppose D contains s% of X and Y, we will consider S% as the supporter of the association rule X≥Y as S%=the number of Transactions X ∪ Y/ The number of Transactions(D); Suppose D contains e% of Y, we will consider e% as the connection of the association rule X≥Y It describes the appearance probability of Y without effect 3.2 The step of data mining association rules In order to discover the meaningful association rules, the association rule mining needs to develop minimum support and confidence level Problems of the association rule can be divided into two parts, and the degree of the support is larger than the minimum support degree Because people pay great attention to the association rule, data mining algorithm has always been used as follows: we need to process the data and seek out all data items to meet the minimum degree of support; it will match the minimum confidence level and then explain R output it 3.3 The usage of algorithm in the system The main objective of TENNIS-DAMS system association rule mining is to analyze the potential association in physical fitness test indicators The core algorithm of the system finds out that we should use the classical association rule mining such as Apriori algorithm and DHP algorithm For example, Apriori is a width-first algorithm, which could be achieved by scanning the database D, and we should only take the same length k (projects contained in each scan trip number) into consideration The algorithm uses the hashing technique DHP to find out the next project, such as Hash function Figure 1. The module of the system & ' "" & " ' (1) 744 9781138028425_Chp_245.indd 744 07/05/15 2:49 PM Hash function considered the first candidate item to generate C1 = ((A), (B), (C), (D), (E)) And then scan the database to get the statistics of the support, thus generate a 1- frequent set L1 = ((A), (B), (C), (D), (E)); In the meanwhile, the statistics can be used to construct the project candidate and so on 3.4 Mining tasks and model evaluation In the system of TENNIS-DAMS, the discovery of the mode can be performed by a combination of the two algorithms In the stage of data preprocessing, we need to divide the attribute of each pattern based on the actual value of the physical indicators, and define the minimum degree of the support and confidence level, by using a DHP algorithm to calculate the data mining And each line of it is an association rules Iilj, the first column is the degree of support, and the second is the confidence level We can get some confirming and enlightening rules of mining from the analysis 3.5 Multi-valued association rule mining Multi-valued association rules and the basic problems have some differences The collection of data items has changed into Ir = I × P × P, i.e., Ir = {(x, l, u) I x∈I, 1∈P, u∈P, 1≤x≤u), (x, 1, u) ∈Ir x represents the attribute values between l and u, I is the set of attributes, P is the set of positive integers The algorithm of the multi-value association rules is MAQA, Boolean association rules problem had changed from multi-value association rules Obtaining valuable rules require the use of Boolean association rules and so on $ $ $ $ $ $ $ $ $ (4) 4.2 Back-propagation training method In the multi-layer network, because of the hidden layer, the output of the network is unable to continue The appearance of back-propagation training will correct the weights the layers It is so called BP algorithm, which had been widely used in networked learning, and it is also used for multi-layer feed forward network learning The main process of it was the back-propagation learning process and forward propagation process The input sample X was added to the layer, and represented the sum of the k, and the function of the neuron excitation is f We can use the following mathematical formula to express the relationship between the variable: (5) The essence is to calculate the minimum value of the error function The BP algorithm is applied to the multilayer feed forward networks Sigmoid function is used as the excitation function, and we can use the following steps to strike the weights (6) The error of each layer dik for the output layer k = m, there are other layers (7) 0 In the light of some learning rule, we usually modified the link between nerve cell weight coefficients w in order to achieve a minimum error function E The learning rule of the neural network is usually divided into the rules related learning, correction rules learning and unsupervised learning 4.1 Overview of neural network technology where w is the weight coefficient of j and f is the nonlinear function of neuron The learning phase of the neural network is a self-improvement, and we apply the following mathematical formula to represent learning formula: 4 Empirical analysis The neural network has obtained from the biological inspiration, connected with others by using a large number of simple processing units and constitute an information system in some way Nonlinear processing, association’s ability and massively parallel computing are all its advantages It can be divided into five parts according to the principles Due to the way of learning, we can study the network by ourselves or learn from the teachers The Hopfield network and BP network of the neural network are widespread use The information processing of it has been divided into the phases of execution and the phases of learning, and the execution stage is to process the neural network such as: 0% $ 1 (3) (2) Right correction factorWij Thresholdθ 745 9781138028425_Chp_245.indd 745 07/05/15 2:49 PM $ $ (8) By (5–25) $ $ (9) Include $ $ $ $ (10) In this process, the output sample X and Y must be performed We should know all the position of the inputs and outputs, the process of the BP algorithm program is shown in Figure 4.3.2 Network model We select 18 items to design the layers of input and output, and evaluate the physical state of the long- distance runners The athlete’s gender, age, height, weight and other 14 indicators of physical fitness will be replaced with nodes The coach usually divided the physical condition of athletes into three states, which is good, generally and poor, and therefore we need to set three different output layers to represent the three different states When we make the choice for hidden layers, it will be extracted the character of the input layer The ability of the neural network could be improved with the increase of the hidden layer, and it will increase the time and date of training By considering the scale and limits on sample data, we could solve the problem by hidden layers, such as BP network model The layers of BP neural network had a great impact on the functional performance of the network, so that the inner layer nodes need to make the right choice With the purpose of evaluating the physical condition, the design of the BP network should be matched with the input nodes and output nodes, training, and the number of samples Since the network is 18 inputs and three outputs, it can be regarded as a characteristic compression process The intermediate layer number N matches 18: N = N: 3, thus available N is approximately equal to After several experiments, we finalized that when the hidden layer nodes is 8, the network performance is relatively stable 4.4 Final results and analysis Table 1. The result of physical test Study sample 50 50 50 50 75 100 125 Figure 2. The process of BP algorithm program 4.3 Evaluation of long-distance runners’ physical condition 4.3.1 Problem statement In the long-distance running training, the coach needs to analyze the athlete’s physical fitness index data carefully in order to evaluate the physical condition On the basis of the data, they divided the physical state into poor, fair and good levels However, the evaluation process is not a simple add, instead of it, it is difficult for coaches to evaluate the process according to their own experiences Test sample Node in hidden layer Rate of average accuracy 100 100 100 100 75 50 25 12 8 8 54.3% 58.24% 56.1% 59.12% 59.4% 61.28% 66.66% Three months is the cycle of the long-distance runner test, so we can predict the period of physical indicators of long-distance runners by adjusting the input samples above network model 4.5 Problems and improvements When we talk about the data entry, there is no denying the fact that problems exist in the data due to the facticity of the data One reason is the small quantity of the data In terms of the network model, we only take 746 9781138028425_Chp_245.indd 746 07/05/15 2:49 PM runners’ physical training is to enhance their overall quality By using the techniques and methods in this paper, the management of physical index and the capability of analysis will be achieved Meanwhile, the relationship between physical fitness and skills testing can be discovered by data mining techniques, and it has provided a scientific and effective method to enhance the ability of athletes Table 2. The result of physical test Study sample Test sample Node in hidden layer Rate of average accuracy 50 100 50 70 20 70 7 48.2% 50.1% 51.3% 100 100 20 20 55.82% 50.1% References the common factors of the distance runners’ physical state into consideration, but ignored their personality factors such as injuries, health conditions and others These factors play an important role in the test, and we should learn to make full use of the information to improve ourselves In the meanwhile, we need to increase the number of the training input samples and filter the data if necessary The sample used in the training can be described as follows: the distribution of three different physical state data is more reasonable, so we need to learn from the coach and personal factors should be considered in the process of input nodes, and we can improve the efficiency of study by accelerating algorithm 5 Conclusion The athletes’ all-round abilities include physical ability and psychology ability which is supplemented for each other The ultimate goal of long-distance [1] Yan Qi, Ren Manying (2012) The training of physical function Journal of the coach in China (01):16–18 [2] Wang Dongyue (2010) The study of the athlete of synchronised swimming Journal of Zhejiang Physical Science, (02):33–36 [3] Isaacs L.S (1998) Comparison of the Vertical and Just Jump system for measuring height of vertical jump for young children Perceptual and Motor Skills, pp:66–77 [4] Huang Tao (2011) The characteristic of the track and field competition Journal of The Technology Information (16):657–658 [5] Teford et al (1989) A simple method for the assessment of general fitness: the tri-level profile Australian Journal of Science and Medicine in Sport, pp:45–65 [6] Jeffreys I (2002) Developing a progressive core stability program Strength Cond J., pp:11–13 [7] Zhou Xingwang (2004) The teaching method of the track and field competition Journal Of Wuhan Physical Institute, (04):125–126 [8] Li Jianguo, Li Jiantao (2011) The research of the core competition for track and field competition Journal of Gunagzhou Physical Institute, (03):78–83 747 9781138028425_Chp_245.indd 747 07/05/15 2:49 PM Electric, Electronic and Control Engineering – Shao, Shu & Tian (Eds) © 2015 Taylor & Francis Group, London, ISBN: 978-1-138-02842-5 An analysis on the maintenance system of electric power equipment based on data mining technology Kangning Wang, Laifu Zhang, Min Jiang, Yun Tian & Tianzheng Wang Electric Power Research Institute, State Grid Shanxi Electric Power Company, Taiyuan Shanxi, China Abstract: The correct evaluation on electric power equipment status has become a key point concerned by related researchers How to extract the rules from multifarious data is also what needs to be solved by data mining technology So there is a corresponding point between the two On the basis of analyzing the application status and process of data mining technology, a status evaluation frame model and a SOA condition based maintenance system of electric power equipment are designed in this paper so as to provide theoretical basis for scientific service of electric power equipment Keywords: electric power equipment; data mining; frame model; SOA; condition based maintenance 1 Introduction foundation for making good strategies for condition based maintenance Zhang Yuan (2008) points out that methods of constructing disaggregated models of data mining now include decision tree, statistics, machine learning, neural network, analogical learning, genetic algorithm, rough set, inference based on cases, etc[1] Data mining is a new technology of data analysis, through which valuable potential information can be extracted from a large amount of data Yang Guoqing, et al (2012) indicates that, with the continued trend of explosive growth of data in electric power system database, it is imperative to introduce data mining technology to the on-line monitoring system of electric power equipment[2] In condition based maintenance system of electric power equipment, quantity of state in the evaluation module is determined through the analysis of basic data based on former experience, which is wasting time and energy When quantity of state is inadequate, reasonable data need to be added according to former experience so as to ensure the accuracy of the final evaluation results Wang Shishuang, et al (2012) proposes that a reasonable method of data processing is urgently needed for making up the insufficiency of existing methods, thereby acquiring more accurate status evaluation information and improving the reliability of condition based maintenance[3] Therefore, in this paper, data mining technology is introduced to the condition based maintenance system of electric power equipment so as to lay a 2 The analysis on the current situation of data mining’s application Chen Chaojin (2009) points out that the expansion of database has brought large amounts of data with the rapid development of information technology, yet information supporting decision-making cannot be identified by people and the need of information mining cannot be met by traditional searching and reporting tools[4] Gao Ningning (2008) puts f orward the concept of data mining, namely the technology of finding potential laws from large amounts of data and extracting useful knowledge[5] Data mining technology has been widely applied in recent years, mainly in such aspects as Telecom customer analysis, agricultural industry data prediction, sales forecasting of retail, web page customization of weblog, fraud of bank customer, biological gene analysis, classification of celestial body, analysis of electric power equipment m aintenance, etc[6] The main research content of this paper is the application of data mining in electric power system There is still a great potential for application of e lectric power system, the one with extremely m ultifarious data information The application situation of data mining technology in electric power system is shown in Figure 749 9781138028425_Chp_246.indd 749 07/05/15 2:50 PM A Application of data mining technology A1 Management informatization A2 Dispatch informatization A3 Dispatch automation A11 Feature extraction of power customer A21 Power system planning A31 Load forecasting A32 Monitoring of equipment operation A33 Feature extraction of power system fault A34 Online safety assessment A35 Statistical analysis on system stability Figure 1. The schematic diagram of application situation of data mining technology in electric power system 3 Description of data mining process The process of data mining is generally composed of five links, namely data selection, data pre-processing, data conversion, data mining and data explanation Besides, such four conditions as effectiveness, innovativeness, serviceability and simplicity need to be satisfied in mining process Data mining process and conditions required are summarized in Table Table 1. Links of data mining process and requirements Links Description Data selection Generally use the Internet to select relevant data with problems needed to be solved Data pre-processing The process of filtering information Data conversion Convert qualitative data into quantitative data, namely feature extraction Data mining Data explanation Requirements Description Effectiveness This requirement implies the importance of laws and knowledge and indicates whether they are applicable to unknown data Innovativeness Search for important models hidden in the database, namely the discovery process of data law Serviceability Evaluate and explain the results of data mining, namely the acquisition of knowledge Simplicity He Youquan (2004) holds that data mining is now being rarely used in electric power departments Applications of data mining in electric power system include status evaluation of electric power equipment, system load forecasting and classification, classification of system operational mode, operation status and equipment monitoring, equipment fault diagnosis, power dispatch optimizing and power system modeling, etc.[7] Instead of being closely related with prior knowledge, it is an important new discovery in the practice process Being useful and interesting for customers Laws shall be as simple as possible and be able to create and explain complex data The main content of this paper is the application of data mining in status evaluation system of electric power equipment, which has a great impact on the condition based maintenance of equipment An excellent evaluation system provides a guarantee for service life and maintenance effect of equipment The flow chart of management on condition based maintenance of electric power equipment and trigger processing of dominant and hidden problems is shown in Figure 750 9781138028425_Chp_246.indd 750 07/05/15 2:50 PM Figure 2. Flow chart of management on condition based maintenance of electric power equipment and trigger processing of dominant and hidden problems The left part of the chart is the management process of condition based maintenance of electric power equipment, the center part is the trigger processing of dominant problems, and the right part is the trigger processing of hidden problems Symbolic meanings of 1–40 are presented in Table Table 2. Symbolic meanings in the flow chart of management on condition based maintenance of electric power equipment and trigger processing of dominant and hidden problems Number Meaning Number Meaning Number Meaning Number Meaning Electric power equipment 11 Testing of agency 21 Maintenance 31 Alarming Assets management 12 data transmission 22 Replacement 32 False alarm recording Online monitoring 13 communication network 23 Monitoring measures 33 Comprehensive confirmation Energized monitoring 14 Manual download 24 Maintenance measures 34 Expert diagnosis Monitoring 15 Data management 25 Review of design 35 Defect elimination Online partial discharge 16 Database 26 Implementing & improving 36 Overhauling Temperature measurement 17 Humancomputer interface 27 Production & operation 37 Technical improvement Oil pressure & air pressure 18 Trend 28 Condition monitoring 38 Releasing Dissolved gas in oil 19 Expert diagnosis 29 Technical section 39 Diagnostic report Gas mass 20 Decision -making 30 Defect confirmation 40 Defect confirmation 10 751 9781138028425_Chp_246.indd 751 07/05/15 2:50 PM