基于开关瞬态高频振荡的变频电机早期绝缘劣化状态监测机理研究

The high-frequency pulse voltages generated by inverter will threaten the long-term operation life of inverter-fed motor insulation and lead to insulation failure. The incipient insulation failure would rapidly develop to serious inter-turn and/or phase-to-ground short-circuit faults, resulting in catastrophic consequences. Actually, insulation failure is a slow process with gradual deterioration of insulation performance. On-line monitoring of insulation deterioration can eliminate the potential fault prior to system breakdown. However, the existing technologies remain great challenges in sensitivity. Previous research indicates that the high frequency resonance impedance of motor winding can effectively reflect the small change of insulation state. The main concept of this project is to monitor the health state of motor insulation using high frequency switching transient oscillation generated by the power electronic inverter itself. By establishing the small time scale transient model of power devices and the high frequency distributed parameter model of motor windings, the interaction of PWM switching transient between inverter and motor windings is explicated. The frequency-response characteristics and resonance mode distribution are obtained, and the formation mechanism of transient oscillation current is revealed. On that basis, the signal acquisition and feature extraction methods of high frequency transient oscillating current are researched, and the quantitative relationship between the variation of insulation state and the response characteristics of oscillating current are clarified. The project research is expected to achieve high sensitivity and non-intrusive on-line monitoring for incipient insulation deterioration of inverter-fed motor.

逆变器产生的高频脉冲电压会威胁变频电机绝缘长期运行寿命，导致绝缘失效并迅速发展为严重的电机匝间和相地故障，引发灾难性后果。实际系统中，电机绝缘失效通常表现为绝缘性能缓慢退化的渐变过程。在线监测电机早期绝缘劣化状态可在故障潜伏性阶段消除隐患，但现有技术在灵敏度上尚存在较大挑战。前期研究发现电机高频谐振阻抗能有效反映绝缘状态的微小变化，为此项目提出利用逆变器自身功率器件高速开关产生的大量高频瞬态振荡电流在线监测电机绝缘运行状态。为深入理解其机理，将建立功率器件小时间尺度瞬态模型和电机绕组高频分布参数模型，明确逆变器PWM开关与电机绕组的瞬态交互作用，获取系统频率响应特性和谐振模态分布，揭示高频振荡的形成机理和特性。在此基础上，研究高频瞬态振荡电流检测与特征提取方法，探索绝缘状态变化对振荡电流响应特征的作用规律和定量关系。项目研究可望实现变频电机早期绝缘劣化状态的高灵敏度和非侵入式在线监测。

变频电机广泛应用于风力发电、新能源汽车、高速列车牵引、船舶电力推进等领域。定子绝缘失效是引发电机故障的重要原因。逆变器高速开关动作产生的高dv/dt冲击会加速定子绕组绝缘老化。在线监测电机早期绝缘劣化状态可在故障潜伏性阶段消除隐患。项目提出利用逆变器自身功率器件高速开关产生的高频瞬态振荡电流在线监测电机绝缘运行状态的研究思路，重点围绕变频电机系统开关振荡机理与特性、绝缘状态劣化对开关瞬态振荡电流的作用规律、绝缘敏感特征提取与状态监测方法展开研究，建立了功率器件开关瞬态模型和电机绕组宽频域模型，揭示了变频电机多模态开关振荡电流的形成机理，明确了各主导模态的传导路径、参与因子和谐振特性。在此基础上，开发了非接触式高频开关振荡电流传感器与检测装置，并提出了基于开关振荡敏感模态和本征特征的电机绝缘早期状态监测方法。研究结果表明：变频电机开关振荡响应电流主要包含高频差模、高频共模和中频共模三个主导模态分量。利用高频开关振荡电流“拖尾分量”能够高灵敏地监测电机绕组早期匝绝缘劣化，利用高频共模谐振频率特征可实现绕组端部绝缘状态的皮法级定量评估。此外，利用多模态开关振荡进行状态感知可以拓宽至变频电机系统其它关键件的状态感知（如功率器件、直流电容、电缆等）。项目共计发表高水平学术论文（SCI/EI）15篇，授权中国发明专利4项，研究成果可为变频电机系统高灵敏状态感知与早期预警提供理论和技术支撑。