Research On Traction Motor Optimization Control For Low Speed Operation Of Locomotive

ABSTRACT:Society and economy development require stronger transportation ability, so locomotive need higher speed and start-up/brake quickly to raise average velocity. Meanwhile, people need ride comfort, which implies locomotive speed should be regulated stably especially during start-up/brake process. There is a conflict between these two demands. Furthermore, energy and environment problem are global important issues, thus efficiency optimization is very meaningful for so large power AC traction motor. To overcome these problems, this thesis researches optimization control strategy of low speed operation and associated technical points.First of all, this thesis analyzes relationship between locomotive operation performance and traction motor control, which is useful for making sure how to optimize motor control. Specifically speaking, ride comfort is mainly determined by traction/braking force curve and their control accuracy; steady deceleration and accurate position shut down depends on braking technique; and motor efficiency optimization is one of the most important methods for low cost operation. Based on the analysis, this thesis discusses key techniques for high performance motor control, sums up existent research and studies remaining problems.Then, this thesis proposes a traction motor optimization controller, which includes three parts:traction force curve optimization, braking force optimization and motor efficiency optimization.(a) Traction force curve optimization:To realize quick and comfort start-up, pulse template polynomial S-curve is used for traction force reference calculation, and locomotive operation resistance force calculation error is corrected.(b) Braking force optimization:Among many kinds of braking force, regenerative electric brake is chosen. Braking force reference calculation is similar to traction force. During braking, motor controller detects running speed, estimates remaining time for shut down, and sends signal to air braking system in ahead for staying brake.(c) Efficiency optimization is realized through assigning copper loss and iron loss by adjusting magnetizing current and torque current in vector control system. And equivalent iron loss resistance estimation is very important for efficiency optimization control.Flux orientation is realized by integrating stator angle frequency in indirect vector control system, but orientation accuracy would deteriorate when traction motor operates in low speed area because of speed measurement sampling frequency decrease. After comparing several kinds of flux model, this thesis proposes an improved voltage model flux observer. The main demerits of voltage model are DC offset in EMF and stator resistance error, because they will result in magnitude and phase error in estimation. Here, a new kind of EMF drift and motor stator resistance error estimator is proposed, with which the errors can be estimated and compensated online.Generally, speed encoder mounted on locomotive has low resolution, so this thesis proposes a consecutive interrupt T-method for speed measurement. Both up-edge and down-edge are captured for speed calculation and every capture generates an interrupt request by setting DSP capture unit with software. Besides, speed calculation uses a smart method to evade encoder pulse duty coherence error. With this method, speed measurement time delay decreases obviously. What's more, this method does not need any hardware modification; it is realized by rational DSP configuration in software. Motor control performance in low speed area is improved evidently with this method.At present, most vehicles mount multi kinds of braking force, and electric braking would be cut off when speed is low, leaving air-braking only until locomotive stops. Air-braking is dependable, but it results in too much human and financial cost, noise and dust. Pure electric brake is very meaningful, while the problem is how to improve motor control performance and zero speed detection during braking down. Based on large inertia and mass of locomotive, traction motor speed could be considered as continuous, so polynomial curve fitting is reasonable. With least-squares polynomial estimation results, both zero speed detection and motor control performance can be improved.At last, this thesis does research on test-bed stability. The analysis is achieved by two steps. First step:assume induction motor is under ideal control, and DC bus stability is analyzed with small signal model of test-bed. Second step:transfer functions between motor variables disturbance and DC bus voltage/current displacement are derived. With frequency characteristic of these transfer functions, test-bed stability is compared with ordinary AC-DC-AC driving system, and oscillation phenomena during experiment have a rational theory explanation.