【作者】 孙伟；

【导师】 徐殿国；

【作者基本信息】 哈尔滨工业大学 ， 电力电子与电力传动， 2016， 博士

【摘要】 由于不需要编码器等转速测量设备,具有无速度传感器矢量控制系统的电力传动系统,具有成本低,可靠性高等优点,在工业领域得到了广泛的应用。随着社会发展,行业对生产效率、成本等要求逐渐提高。从生产成本、安装空间、可靠性、安全性方面考虑,需要变频器具有直驱性能,摒弃价格昂贵、体积庞大的降速齿轮箱、升速齿轮箱,摒弃编码器、旋转变压器等速度传感器,这就需要变频器具备极高的调速比,可以在极低转速运行方面满足工业场合的需求。但是目前感应电机无速度传感器矢量控制技术在电机极低速运行方面仍然存在以下几个问题:(1)估计转速在低速发电区域的不稳定性;(2)在低速区域估计转速和估计转子磁链对电机参数的变化很敏感;(3)低速运行时的逆变器非线性误差影响转速估计精度;(4)当定子电流频率为0Hz时,转子转速不可观测;(5)矢量控制电流环带宽较低导致低速突加负载时转矩响应较慢。本文将分别针对以上五个问题,对应给出相应的解决方案。本文分析了现阶段全阶磁链观测器误差反馈矩阵的四种主流设计方法。在观测器收敛速率,估计转速稳定区域,估计转子磁链对电机参数鲁棒性和估计转子转速对电机参数鲁棒性四个方面,将四种误差反馈矩阵进行了对比分析。针对估计转速在低速发电区域的不稳定问题,本文建立全阶磁链观测器转子磁链观测和转速观测数学模型,推导出估计转速传递函数,分析得到造成估计转速在低速发电不稳定问题的原因,即估计转速开环传递函数在低速发电区域存在具有正实部的零点,当估计转速增益足够大的时候,估计转速闭环传递函数极点具有正实部,系统发散。针对这个问题,设计了基于劳斯稳定判据的全阶磁链观测器误差反馈矩阵并进行了改进。改进后的误差反馈矩阵,不但能够保证估计转速在低速发电区域的稳定性,而且能够提升估计转子磁链和估计转速对电机参数的鲁棒性。针对估计转速在低速运行时对电机参数很敏感的问题,在传统转速估计自适应律中引入观测磁链误差,并基于改进的转速估计自适应律,提出了两种不同的提升估计转速对电机参数鲁棒性的设计方案。第一种利用现有误差反馈矩阵,改进其估计转速自适应律,设计估计转速自适应律系数以保证估计转速在低速发电区域的稳定性。最后得到传统反馈误差矩阵配合改进转速自适应律,与传统组合方法相比,估计转速对电机参数的鲁棒性得到了明显的提升;第二种为采用改进的估计转速自适应律,设计误差反馈矩阵。与传统方法和第一种方法相比,第二种方法得到的观测器很大程度提升了对电机参数鲁棒性。针对逆变器非线性误差电压问题,建立考虑逆变器非线性误差电压的估计转速传递函数,量化分析了逆变器非线性误差对估计转速的影响,并通过以下两个方面,减弱和抵消逆变器非线性误差对估计转速的影响:(1)设计误差反馈矩阵,减弱逆变器非线性误差对估计转速的影响;(2)提出逆变器非线性误差自适应辨识策略,离线对逆变器非线性误差进行辨识,并当电机运行时在线补偿。针对定子电流频率为0Hz时电机转速不可观测的问题,本文分析了转速不可观测的原因,在此基础上,提出了利用全阶磁链观测器输入误差电压来估计转子转速的方法。当定子电流频率为0Hz时,传统转速估计方法无法估计转速,此时由观测器误差电压进行转速估计。文中分析了该方法的稳定性,并通过实验验证了该方法能够在定子电流频率为0Hz时保证电机能够长时间带载稳定运行,并且具有很强的对电机参数鲁棒性。针对矢量控制电流环带宽较低导致低速突加负载时转矩响应较慢的问题,本文采用了预测控制替换传统PI控制,但预测控制效果对电机参数很敏感,本文提出了一种鲁棒性预测控制方法,在提升系统带宽的前提下保证控制器具有较强的鲁棒性。文中分析了鲁棒系数与电流环带宽以及系统鲁棒性的关系,最后通过实验验证了所提方法的有效性。更多还原

【Abstract】 In applications speed sensor is expensive and unreliable in hostile environments.It is necessary to use speed sensorless drives for instead.In general the advantages of speed sensorless drives include low cost,reduced hardware complexity and size,improved reliability and less maintenance requirements.The main disadvantage of speed sensorless drives is the poor performance when induction motor runs at low speed range or at zero speed point.The reasons mainly include:(1)there is the unstable range when the motor operates at low speed regenerating mode;(2)the estimated flux and rotor speed accuracy is sensitive to the motor parameter deviation;(3)the inverter non-linear error has large influence on the flux and speed estimation at low speed range;(4)the rotor speed can not be estimated when the current frequency is zero;(5)the response of torque is slow at low speed range because the bandwidth of current loop is limited.In this paper,all above five problems at low speed range will be solved.Many approaches of speed sensorless induction motor drives have been proposed in past years.There are mainly two basic approaches in sensorless control system.One approach is based on fundamental model of induction motor.Fundamental model has poor performance when motor runs at low speed range in motoring mode or in regenerating mode.The other approach is signal-injection-based method for speed estimation,which is based on rotor slot harmonics,main inductance saturation,artificial saliency or rotor slot leakage.Wit h signal injection,the motor can run at zero speed.In this paper,the speed sensorless induction motor drive is researched for the low speed operation:The performance of adaptive full order observer for sensorless induction motor drives can be improved further though it has been researched for many years.In this paper,two different design methods for improving the speed estimation stability and robustness are proposed.1)In classical speed estimation algorithm used in adaptive full order observer,the estimated rotor flux error is neglected.A robust design method of feedback gains with classical speed estimation algorithm is proposed.Compared with existed three main design methods of feedback gains,the stabilit y of speed estimation and estimated speed error are improved.2)A robust speed estimation algorithm with estimated rotor flux error and the corresponding feedback gains are proposed in this paper.Through Lyapunov’s theorem,the stability of proposed speed estimation algorithm is proved.The stability of flux estimation and estimated speed error can be improved much more.Finally the feasibility of two proposed methods is verified by experiments.The performance of speed sensorless induction motor drive is poor at low speed range.That is because the inverter non-linear error,such as inaccurate motor parameters,deadtime,turn on/off delay and on-state voltage drop of power switches,has strong influence on the speed estimation at low speed range.In this paper,the feedback gains of adaptive full order observer(AFO)are designed to guarantee the stability of speed estimation and to reduce the influence of inverter non-linear error on the speed estimation.In addition,the estimated current error caused by inverter non-linear error is identified off-line.It is stored in the look-up-table to compensate the influence of inverter non-linear error on the speed estimation.The feasibility o f proposed method is verified by experiments.It is well known that the operation of speed sensorless induction motor dr ive at zero stator current frequency is unstable theoretically if signal injection is not used.In this paper,the unstable reason is analyzed.The error between stator input voltage of adaptive full order observer and motor is introduced for the operation at zero stator current frequency without any signal injection.When the stator current frequency is reduced to zero,the traditional speed estimation fails to estimate the rotor speed but the rotor speed can be estimated by the proposed stator input volta ge error.The stability of the proposed method is analyzed.Though the unstable problem at zero stator current frequency can be solved by the proposed method,there is a disadvantage that an error between estimated and actual rotor speed is caused.Finally the feasibility of proposed method is verified by experiments.A novel robust predictive control scheme,with indirect vector control system,for the induction motor is proposed in this paper.In order to improve the current loop bandwidth in control system,the restrictions to current-loop bandwidth improvement is analyzed.The predictive control scheme based on mathematical model of induction motor in synchronous frame,is proposed to improve the current-loop bandwidth and the performance of dynamic response.Meanwhile reasons of destabilization in system are analyzed and two robust factors are added to predictive current controller to upgrade the robustness of system and reduce the sensitivity to motor parameters.Experiment results show that the propose d robust predictive current control scheme improves the robustness of control system,and raise system response time.更多还原