Research On Speed-Sensorless Control Strategy For Traction Drive System Of Medium-Low Speed Maglev

Traffic congestion acts as an important role in restricting urban development.It is an effective solution to vigorously develop new-generation rail transit systems.As one of the promising candidates of the new-generation rail transit systems,medium-low speed maglev system has been booming in recent years due to its comparative advantages,e.g.,small turning radius,strong climbing ability,cheap maintenance cost,etc.In a medium-low speed maglev vehicle,the linear induction motor(LIM)drive system is generally employed as the power source.The reliable operation of the LIM drive system is of high significance for the medium-low speed maglev system.A development trend towards ensuring the reliable operation of the LIM drives is to adopt high-performance speed-sensorless techniques.The procedure involves speed information detection by using various speed estimation schemes,and accordingly,removing speed sensors.In this paper,the LIM drives applied in medium-low speed maglev is focused on,and high-performance speed-sensorless techniques suited for LIM drives are presented.More details are shown as follows.In Section II,starting from the analysis of the operation principle of the LIM,the mathematical model of the LIM is first given.Afterward,the vector-controlled LIM drive system is realized with the vector control theory.During the operation of the LIM drives,the dynamic end effects may occur.As a consequence,the parameters of the LIM(e.g.,magnetizing inductance)vary obviously due to the dynamic end effects,which may further result in the performance degradation of vector-controlled LIM drives.To address this,a parameter on-line identification scheme using the adaptive full-order observer(AFO)is proposed in Section II.In this scheme,the state-space model of the LIM is first deduced according to the mathematical model of the LIM,where the primary current and secondary flux are acted as the state variables.Based on this,the AFO model of the LIM is accordingly obtained.A model reference adaptive system(MRAS)for parameter on-line identification is then designed.The state-space model and the AFO model are used as the reference model and the adaptive model,respectively.The Popov’s hyperstability theory is then selected as the adaptive law for the MRAS-type scheme.Eventually,the magnetizing inductance on-line identification is achieved.A speed estimation scheme based on the second-order sliding mode observer is presented in Section III.In this scheme,according to the super twisting algorithm(STA),the state-space model of the LIM is rearranged.By doing so,the speed information is absent in the rearranged model,and accordingly,this model is taken as the reference model for speed estimation.Meanwhile,the current model is used as the adaptive model.Consequently,the MRAS-type speed estimator is achieved.Moreover,the stability analysis of the proposed speed estimation scheme is performed through the Lyapunov stability theory.In practice,the primary resistance variations of the LIM may adversely affect the speed estimation performance.Thus,a parallel primary resistance on-line identification scheme is introduced into the proposed speed estimation scheme to enhance the robustness against parameter variations,and hence,guarantee the performance.As for the model-based speed-sensorless techniques,the phase-locked loop(PLL)-based speed estimation schemes attract much popularity due to their advantages,e.g.,easy implementation,high flexibility,acceptable performance,etc.However,the conventional PLL-based schemes may suffer from performance degradation during acceleration and deceleration processes.In order to achieve accurate estimation,attempts of the new-type PLL-based schemes are made in Section IV,including the type-3 PLL scheme and the dual-loop PLL scheme.In Section IV,a brief introduction of the conventional PLL-based estimation scheme and the performance analysis are first presented.Then,the properties of the type-3 PLL scheme and the dual-loop PLL scheme are detailed.In the type-3 PLL scheme,a new loop filter(LF)is designed to replace the proportional-integral(PI)controller-based LF in the conventional PLL scheme.This scheme can be accepted as a type-3 control system that can perform estimation with a satisfactory accuracy during acceleration and deceleration processe.By comparison,in the dual-loop PLL scheme,extra loops are introduced on the basis of the conventional PLL-based scheme.By doing so,a performance enhancement during acceleration and deceleration processes is made.Furthermore,due to the limited disturbance elimination capability,the new-type PLL-based schemes cannot successfully deal with the disturbances in LIM drives(e.g.,harmonics,parameter variations,DC offsets,etc).Considering this,the adaptive filters and the closed-loop flux observer are equipped into the implementation of the new-type PLL-based schemes to mitigate the adverse effect of disturbances.More specifically,the adaptive filters are responsible for mitigating the harmonics and the closed-loop flux observer is used to attenuate the parameter variations and DC offsets.Frequency-locked loops(FLLs)are regarded as the emerging synchronization techniques in grid-connected applications,and accordingly,gain considerable attention.The second-order generalized integrator-frequency-locked loop(SOGI-FLL)and reduced-order generalized integrator-frequency-locked loop(ROGI-FLL)are two notable examples.Due to their frequency estimation units,the SOGI-FLL-and the ROGI-FLL-based estimation schemes still achieve unsatisfactory performance during acceleration and deceleration processes for motor drive system applications.Additionally,despite being effective in ensuring acceptable accuracy,the new-type PLL schemes typically bring concerns in terms of system dynamics and stability margin.Considering this,the new-type FLL-based estimation schemes are explored for speed-sensorless control of LIM drives in Section V.A thorough review of the conventional SOGI-FLL-and ROGI-FLL-based estimation schemes along with the performance analysis is first provided.Then,with the assistance of the simplified model,the performance of the new-type PLL scheme is evaluated.On the basis of the SOGI-FLL-and ROGI-FLL-based estimation schemes,the enhanced FLL-based schemes(including the enhanced SOGI-FLL-and enhanced ROGI-FLL-based schemes)are elaborately designed.Unfortunately,according to the performance analysis of these schemes,it is indicated that the enhanced FLL-based schemes perform accurate estimation at the cost of increasing the system order.Consequently,the stability margin is undesirable.In light of the above,several adaptive FLL-based estimation schemes are developed for speed-sensorless control of LIM drives,in which accurate speed estimation is achieved through designing new frequency estimation units,rather than solely increasing the system order.Benefiting from this,the adverse effects on the stability margin are effectively mitigated.However,in the adaptive FLL-based estimation schemes,the estimated speed should be fed back,which may introduce new concerns.Therefore,the tracking differentiator(TD)-FLL(TD-FLL)-based scheme is proposed to handle such scenarios.In this scheme,the estimated speed feedback is cancelled due to the use of the TD.This can further maintain the estimation performance.