双馈异步发电机内部故障的振动(声学)机理分析与机电(声)融合诊断研究

The doubly-fed induction generators (DFIG) normally operate in severe environment with high failure rate, which will cause a serious influences duo to the failures are difficult to be detected. Therefore, this problem will become more and more prominent with the increasing of the new units, especially with the increasing operations life of the active units. Based on the forthcoming concluding fund (51177039) considering the electrical characteristics of DFIG common faults, this project will further study the electrical characteristics of DFIG electrical faults (short circuit of stator and rotor windings, strand broken, etc.) and mechanical failures (dynamic eccentricity, static eccentricity, etc.) based on the theory of vibration (acoustic). By using the multi-loop method, analytical method and finite element method, these faults will be analyzed in detail, respectively. By setting different faults, the changes of electromagnetic field distribution, the electromagnetic force and the vibration between normal and fault conditions will be considered, then the vibration characteristics of DFIG failures are explored. Through developing different forms of fault simulation motors, some experimental research will be performed. Some necessary comparing results of the four methods are developed, then the reasons of the differences will be analyzed carefully. Furthermore, the main sources of vibration affecting the running wind turbines, including the natural vibration on normal state and the new vibration on the fault conditions, will be studied. Finally, combing with the results of the fund (51177039), the fusion diagnosis of the electrical and mechanical (sound) characteristics of DFIG internal faults will be studied thoroughly.

双馈异步发电机（DFIG）运行环境恶劣，故障率高，故障难以发现，影响大。随着新机组投运量进一步扩张，特别是现役机组运行年限增加，问题将更加突出。本课题在即将结题的基金(51177039)对DFIG故障电气特征研究基础上，进一步提出基于振动（声学）对DFIG的电气故障（各种短路、断股）、机械故障（静、动态偏心等）的故障机理进行研究。采用多回路法、解析法和有限元对上述故障进行并列分析；设置不同故障，研究DFIG正常状态与各种故障形态的电磁场分布、电磁力和振动变化，进而研究DFIG故障的振动特征；研制故障模拟电机，并进行故障模拟实验研究；将四种研究方法的研究结果进行对比，分析差异；对风电机组运行过程中影响电机的各振动源进行研究，包括正常状态的固有振动、故障状态的新增振动；进行DFIG故障特征识别；最后结合51177039的研究成果及其他信息，进行DFIG内部故障的机电（声）融合诊断研究。

通过对双馈异步发电机故障4年的理论与实验研究，本课题主要取得下列研究成果：.（1）定子绕组内部故障电气量方面，成果：.1）电流的变化，结论：定子绕组内部故障（绕组匝间短路）与定子绕组外部故障（电网电压不平衡等）均使定子侧线电流不对称，但其波形有本质的区别：定子绕组匝间短路时，定子侧线电流的幅值不变；电网电压不平衡时，定子侧线电流的幅值呈周期性变化。2）电磁转矩的变化，结论：匝间短路故障时电磁转矩的波形是一波动的曲线，且匝间短路故障越严重，其波动范围越大。对电磁转矩进行频谱分析发现，故障后其直流分量值变小。3）平均瞬时功率变化，结论：故障后其时域图是一条曲线，且定子匝间短路故障越严重，其时域谱的波动范围越大，其直流分量值越小，出现特殊频率成份（公式在此网页文本中不能显示） 。.（2）定子绕组匝间短路后的振动变化，结论：正常运行时振动信号的100Hz频率幅值很小，定子绕组匝间短路故障后100Hz频率幅值变化很大；定子绕组匝间短路故障后振动信号能量从500-1500Hz频段向0-500Hz频段转移，且随着故障程度的增加，能量转移的越多。.（3）电刷滑环系统的烧伤故障，结论：将励磁电流中高于100Hz的周期性高频分量作为故障特征量，并进一步将其能量熵值用于诊断电刷滑环系统的烧伤程度。 .（4）为诊断双馈风力发电机转子绕组的过热故障，提出了一种新的参数在线辨识诊断方法。通过数学模型辨识转子绕组电阻和温度值,为故障初期的监测预警和进一步的措施维护提供了数值参考。提出了基于转子变换器网侧励磁电流中特定频率的双馈风力发电机转子绕组的不平衡故障的诊断方法。.（5）基于振动的转子绕组的不平衡故障研究方面，提出基于振动信号轨线的故障诊断方法。结论：风机正常运行时，相平面的振动信号轨线等效为一个空心的畸变椭圆且沿主对角线y=x闭合；随着故障程度的加深，相平面振动信号轨线脱离主对角线且畸形椭圆不断地内缩，最终形成一个带毛刺的实心球团。这种定性的图像分析为风机转子绕组不平衡故障的检测与融合提供了新思路。