An Expert System Approach For Analyzing And Evaluating The Testing Results Of CTCS-3 On-board Equipment

The CTCS-3 train control system, which has been widely deployed in China Railway High-speed (CRH for short) trains with speed over 300km/h, is the key and highly efficient system that guarantees the safety of trains. As the most important component of the train control system, it is necessary to carry out a series of testing and analyze testing results on the on-board equipment to verify that the on-board equipment complies with the system requirement specifications. In the current testing environment, the testing results acquisition, analysis and evaluation of testing results are manually applied, which is time-consuming, with low efficiency and high demand for manpower. Furthermore, it can probably cause safety risk if any inappropriate analysis is made by human. Therefore, it is significantly important to carry out the research on the methodology for analyzing and evaluating of testing results based on expert system.In this thesis, the methodology for analyzing and evaluating testing results is proposed based on the CTCS-3 on-board equipment, a laboratory testing platform in order to improve the efficiency and accuracy. Besides, a relevant software tool is implemented at the same time. The main contribution of this thesis is listed as follows.1. To address the disadvantages of manual analysis on testing results in current testing environment, the overall design of the system and the frame of testing results analysis and evaluation are proposed based on expert system.2. Studying technical specifications and field testing experience of the on-board equipment, the necessary knowledge related to analysis and evaluation of the testing results is extracted by means of direct acquisition method. Subsequentially, a knowledge tree is applied to collate the acquired knowledge, following by rules design and presentation based on production representation. Moreover, the knowledge base of the expert system and the basic testing results database is conducted with relational database.3. The testing results analysis and evaluation method based on the rules is designed, including several key issues such as searching for the rule strategy, the comparison strategy and the method of results evaluation. Firstly, the testing results are acquired by Driver Machine Interface (DMI) and Data Recorder, and then preprocessed and stored in the database. Next, the required rules are selected from rule base according to the different triggers and the rule set is built up. The rules in rule set are then simplified to a single sentence consisting of a condition and a conclusion, followed by comparison between the regularized data and the testing results. Finally, the analysis results are evaluated and the analysis and evaluation report is generated.4. An analysis and evaluation software tool, for testing on-board equipment, based on expert system is designed and implemented, under Visual Studio 2010 SDK and the Access database. Within this tool, the maintenance and management for the knowledge base of the expert system and the basic testing results database is implemented as well. Meanwhile, the software tool is verified by analyzing the testing results of "Zaozhuang west station to Xuzhou east station of Beijing-to-Shanghai Railway".