Expert System Applied to High-Speed Railway Track Circuit Coding and Its Simulation Yan Li, Dewang Chen, and Xuan Yang State Key Laboratory of Rail Traffic Control and Safety, Beijing Jiaotong University, Beijing, China {09120301,dwchen,10120337}@bjtu.edu.cn Abstract. Track circuit coding can be realized by Train Control Center in Highspeed Railway to keep the train running safe. This paper looks into the track circuit coding method using expert system, in which the reasoning process of track circuit coding was realized using software simulation with the data of Beijing-Shanghai High Speed Railway. Combined with test cases, the joint debugging is realized and the track circuit coding expert system is assessed. Final simulation results support that this prospecting method is both feasible and effective, which can secure the safety of the track circuit coding information as well. Keywords: High-speed Railway, Expert System, Track Circuit Coding, Simulation. 1 Introduction With the gradual opening and running of Beijing-Tianjin, Wuhan-Guangzhou, Zhengzhou-Xian, Shanghai-Nanjing, Beijing-Shanghai High-speed Railway, the China High-speed Railway Passenger Line is now leading the development of the world high-speed railway using the CTCS-3 Train Control System, achieving the highest running speed up to 350 km/h [1]. As an important part of CTCS-3 Train Control System ground equipments, Train Control Center (TCC) produces the permit order for the train operation to realize the automatic control of the train tracking interval and the safe operation of the train speed [2]. While in China, relatively there aren t too many researches on the track circuit coding algorithm, except those that mainly focus on how to build the track circuit coding mathematical model and simulation realization[3][4]. In fact, the track circuit coding varies under different circumstances. Line type (up-line or downline) as well as direction (the forward or the reverse direction) and so on should be considered in the coding in section. Besides, route type needs to be considered in the coding in station. And it will get more complicated when there are more trains. If not all the circumstances have been taken into fully consideration, the safe running of the H. Deng et al. (Eds.): AICI 2011, Part I, LNAI 7002, pp. 517 524, 2011. Springer-Verlag Berlin Heidelberg 2011
518 Y. Li, D. Chen, and X. Yang train will be impacted. Thus, it is quite necessary to build one database with all information included. The expert system contains large amount of knowledge and experiences of expert in one field, therefore, the expert system can be applied to studying the track circuit coding, which helps eliminate problems in TCC as mentioned above (e.g. complicated system, various circumstances), meanwhile, the safety and reliability can be assured as well. In addition, as the train fundamental plan is relatively stable and the line data (e.g. the block section) is relatively fixed, some common circumstances can be used as case studies and saved in the knowledge database. Then the expert system can be applied to search matched case among these common ones conveniently without complicate coding process so as to improve the efficiency of the track circuit coding. This paper introduces the track circuit frequency in high-speed railway, applies the expert system to study the coding method, simulates and realizes the track circuit coding expert system. Combined with the test cases, the function of the track circuit coding expert system is verified. 2 Track Circuit Frequency The interval track circuit applies the ZPW-2000 (UM) series jointless track circuit controlled by the computer coding. The normal coding sequence of the track circuit is: L5-L4-L3-L2-L-LU-U-HU, which can meet the demand of safe running for the 300 km/h train in CTCS-3 train control system [2]. The track circuit coding frequency consists of low frequency and loading frequency, in which the loading frequency of each track circuit is a fixed value. The low frequency is mainly used to reveal the coding frequency to get the different speed limitation information [5].This information is calculated and sent to the trackside equipment by TCC. Part of the low frequency coding is defined as: L code: the train is allowed to run according to the specified speed. U code: the train needs to be slowed down to the specified speed level. L5 code: the train is allowed to run according to the track speed, when at least 7 block sections in the front are free. LU code: the train is allowed to run watchfully according to the track speed. HU code: the train is required to stop in time. UU code: the train speed is limited (default speed: 45 km/h) and the ground signal (the train is approaching) open the route through the switch in side line [6]. 3 Track Circuit Coding Expert System The expert system is a computer software system which contains a large number of authority knowledge, where knowledge can be applied to solve the time problem in the specific region through reasoning and analysis process [7].
Expert System Applied to High-Speed Railway Track Circuit Coding 519 3.1 Overall System Structure The overall system structure of track circuit coding of TCC is as shown in Figure1 below, including line database, real-time database, data processing, overall database, reasoning machine, knowledge database, output section and so forth. Fig. 1. Overall System Structure The main function of each part is as follows: 1) Line database. It can save line data, including station files, route information, track circuit information, balise information and so forth. 2) Real-time database. It is used to save all kinds of real-time data, including simulation order, route information from Computer Based Interlocking (CBI), and track circuit occupation information from trackside and so forth. 3) Data processing module. It is used to process data in the real-time database, and abstract the needed information. 4) Overall database. All kinds of data (e.g. real-time data, intermediate results in the reasoning process) are saved in the overall database. It acts as a monitor in the reasoning process. 5) Knowledge database. It is mainly used to save the specific knowledge provided by the region experts, including all kinds of rules and case library for the track circuit coding. 6) Reasoning machine. It can be able to select relevant knowledge using certain reasoning strategies from the knowledge database, and then realize the reasoning according to the context in the overall database. 7) Output module. It is used to send the track circuit coding information to the other relevant equipments and reveal that in the interface.
520 Y. Li, D. Chen, and X. Yang 3.2 Building of the Knowledge Base In this paper, the knowledge database consists of rule database and case library. The rule database is used to save the track circuit coding regulation, and the case library is used to save different cases where lines, routes or locations vary, therefore, it is convenient to extend the knowledge case. Fig. 2. Knowledge Base Figure Rule Database and its Knowledge Representation The cab signal low frequency information distribution and its corresponding fundamental signal representation is defined in the railway industry standard in China- Cab signaling message define and allocation(tb/t3060-2002)[6], This paper summarizes the track circuit coding principles and builds rule database according to the standard. 1) Rule Database The track circuit coding has its corresponding regulation, so regulation representation method can be used to represent the knowledge in knowledge database.
Expert System Applied to High-Speed Railway Track Circuit Coding 521 According to IF-THEN rules, the object-orientation knowledge based on the regulation can be obtained. rule: // rule 1 If interval section and tracking // the premise then HU-U-LU-L-L2-L3-L4-L5 // the conclusion rule: // rule 4 If interval section and arrival side-line route and speed limitation at side-line is 45 km/h // premise then UU-U2-LU-L-L2-L3-L4-L5 // conclusion rule: // rule 7 If in station and no route // premise then code the switch B, code the track HU, code the route protection section B // conclusion rule: // rule 9 If in station and departure route and side-line 45km/h //premise then code the track UU, code the switch area according to departure section, the departure section applies interval section tracking coding // conclusion The other specific rules are as shown in Figure 2. 2) Knowledge Representation The knowledge can be represented by the regulation; the data structure under Visual C++ mode is as follows: STRUCT RULE_{ char*result; // regulation char code; // the track circuit coding type rule* next; // point to the next algorithm }; Case Library The track circuit coding has many circumstances. Different matters should be considered in the coding in section, such as whether it is up-line or down-line, forward or reverse direction, in rear or advance of station and so on, while the track circuit coding in station need to consider route type and so forth. It ll become more complicated with multi-trains. This paper creates the case library for different circumstances. When new circumstance (unconsidered) comes out, reason the conclusion using rule database. And, at last this new conclusion will be added into case library, so case library can be extended.
522 Y. Li, D. Chen, and X. Yang Construction of the Reasoning Machine The reasoning machine is responsible to analyze and reason the needed data and make a conclusion at last[8]. Forward reasoning is selected in this system, and the reason process is as follows. Fig. 3. Reasoning Machine Structure 4 Simulation and Testing 4.1 Simulation Beijing-Shanghai High-speed Railway data are used in the simulation. It adopts CTCS-3 Train Control System whose ground equipment consists of Radio Block Center (RBC), Computer Based Interlocking (CBI), Train Control Center (TCC), Centralized Train Control (CTC), and trackside equipment (including balise and track circuit) and so forth. TCC simulator interface is as shown in Figure 4, with the station module, the section module, the train control center information module, and the station coding sequence module. After receiving simulation starting order, information in line database is read, and the program will produce TCC simulator automatically.
Expert System Applied to High-Speed Railway Track Circuit Coding 523 Fig. 4. TCC Simulator Interface 4.2 The Testing Result and Analysis The joint debugging is based on CTCS-3 Level train control system simulation testing platform. Data in debugging cases are collected from Xuzhou East to Suzhou East in Beijing-Shanghai High-speed Railway. Then functional testing is done repeatedly according to the test cases. Test cases used to test the function of track circuit coding expert system are list in Table 1. TCC simulation system based on the expert system is assessed through repeated tests on different cases with an accuracy rate of 100%. Table 1. Examples of Test Cases Test Cases the departure, stop and passing by testing through the track line the departure, stop and passing by testing through the track line the route guiding testing for the arrival and dispatch the bent in and direct out, the reverse running testing the direct in and bent out, the reverse running testing the reverse bent in and direct out, the forward running testing the reverse direct in and bent out, the forward running testing Testing Results Other block sections:u-lu-l-l2-l3-l4-l5 Track line:hb Other block sections:u-lu-l-l2-l3-l4-l5 The simulation results show that track circuit coding method based on the expert system can get track circuit low frequency information correctly, and provide effective promise for the safety of cab signal of high-speed train.
524 Y. Li, D. Chen, and X. Yang 5 Summary This paper mainly focuses on the track circuit coding in high-speed railway. In order to get track circuit frequency correctly, the knowledge and reasoning mechanism of expert system is applied for track circuit coding. Using data from Xuzhou East to Suzhou East in Beijing-Shanghai High-speed Railway, track circuit coding expert system is developed, and some test cases are designed to evaluate validity of the system. The results show that the track circuit coding expert system can get track circuit frequency correctly and efficiently, and can keep the safe of cab signal. Acknowledgments. This research work was supported by Beijing Nova Program (2010B015), the Fundamental Research Funds for the Central Universities (2011JBM157), and Independent Research Project of the State Key Laboratory of Rail Traffic Control and Safety (RCS2008ZZ001 and 2009ZT004). References 1. Xue, L.: The CTCS-3 Simulation and Testing Platform Research on Multi-Train Simulation Subsystem. Beijing Jiaotong University, Beijing (2009) 2. Zhang, S.: The technology plan of CTCS-3 Simulation Platform, pp. 122 123. China Railway Publishing House, Beijing (2008) 3. Li, M.: The research on function simulation of the CTCS-2train control centre for Passenger Dedicated Line. Southwest Jiaotong University, ChengDu (2009) 4. Li, J.: The train control subsystem simulation design and research for the CTCS-2 train control centre. Southwest Jiaotong University, ChengDu (2010) 5. Zhao, L., Guo, J., Li, H., et al.: The simulation analysis of influence on jointless track circuit signal transmission from compensation capacitor based on transmission-line theory. In: 2009 3rd IEEE International Symposium on Microwave, Antenna, Propagation and EMC Technologies for Wireless Communications, vol. 9(11), pp. 1113 1118 (2003) 6. Ministry of Railways. TB-T3060-2002.Cab Signaling Message Define and Allocation. The People s Republic of China Ministry of Railways, Beijing (2002) 7. Tang, P.: The Expert System Lecture. BeiJing KeHai Training Center, Beijing (1987) 8. Sun, Y., Zhang, J., Huo, K., Fu, Q.: Application of Simulation Technology Based on the Support of Expert System in the Testing System of Locomotive. Basic Automation 10(3), 242 244 (2003)