Research On Real-Time Detection And Calculation Algorithm Of Locomotive Traction Performance

Traction performance as an important technical characteristic of electric locomotive and train-set is not only the key parameter of train operation and control, but also the reference of new train design and used train transformation. Under the circumstances of putting new railway lines into use or improving passenger train speed or increasing freight train tonnage or reforming driving system, locomotive traction performance test has become a foundational technical job. An improved algorithm for real-time calculation of traction force is proposed in this paper. Through the modular designing of hardware and real-time detection of locomotive running status to achieve the purpose of calculating locomotive traction force accurately.Based on the analysis of current different methods of measuring traction force, a real-time calculation sub-model algorithm of traction force combined with dynamic test vehicle data according to "train traction calculation regulation" as the theoretical basis and SS3B electric locomotive as study object is established in this paper. The start-up phase traction force algorithm model is determined using multiple linear regression and curve fitting methods. The operational phase traction force algorithm model is determined using correlation analysis with the introduction of generalized efficiency. Locomotive traction force can be calculated accurately by real-time detection of the traction motor voltage, current, velocity and the introduction of test line data.Hardware platform of locomotive running status including analog signals,switching signals, RS485 bus signals is introduced in detail. More detaileddescription of the design for hardware circuit is given. Aiming at real-time andaccurate requirements of the system, key issues of hardware design disposal methodsare given.In this paper, interactive PC interface is built by Lab Windows / CVI to realize data acquisition, storage, playback and analysis functions. The simulation test of the algorithm model shows the result accuracy is improved and achieves the desired goal. It is proved that the design is reasonable.In the end, the works and achievements in this paper are summarized.