【作者】 王旭；

【导师】 王昕；

【作者基本信息】 上海交通大学 ， 电气工程（专业学位）， 2019， 硕士

【摘要】 变压器油在油浸式电力变压器中发挥着绝缘、散热冷却、消弧等重要作用,由于杂质污染、变压器油老化等导致变压器油性能下降,给电力变压器的正常运行造成隐患。但现有的变压器油检测方法检测过程繁琐、获取的油性能参数单一、检测效率低。因此,本文提出了基于多频超声波技术的电力变压器油检测方法,并设计制作了电力变压器油参数检测系统,从而实现对多个变压器油参数的快速、准确检测。首先,本文研究了变压器油参数的多频超声波检测技术原理。通过分析超声波检测的原理,得到超声波检测的三种常用方法。通过研究多频超声波在变压器油中传播时的波动特性和衰减特性,得到超声波声学参量与变压器油参数之间的关联,总结了变压器油参数的多频超声波检测技术原理。其次,在研究多频超声波检测方法的基础上,设计并制作了变压器油参数的多频超声波检测硬件系统。首先对超声波声速和衰减系数的测量方法进行了研究,确定了硬件系统的设计思路。接着对超声波换能器和数据采集卡进行了选型,并设计了多频超声波发射驱动电路和回波处理电路,最终形成了变压器油参数的多频超声波检测硬件系统。该系统可以实现不同频率超声波的发射和回波的采集,可获取超声波的声速和衰减系数等参数,将检测数据传输到上位机软件,从而实现变压器油参数的高效、准确检测。最后,设计并开发出变压器油参数检测软件系统。首先利用改进灰狼算法优化的BP神经网络训练样本,建立变压器油参数的多频超声波检测模型。以检测模型为基础,设计并开发了变压器油参数检测软件系统,可以对多频超声波检测数据进行处理,快速生成变压器参数的检测结果,并对检测结果进行分析,给出变压器油的运行状态。系统实现了对变压器油参数的高效、准确检测,能够对实际检修工作起到很好的指导作用。更多还原

【Abstract】 Transformer oil plays an important role in insulation,heat dissipation and arc suppression in oil-immersed power transformers.The performance of transformer oil decreases due to impurity pollution and aging of transformer oil,which causes hidden dangers to the normal operation of power transformer.However,the existing transformer oil detection methods have the disadvantages of tedious detection process,single oil performance parameters and low detection efficiency.Therefore,a power transformer oil detection method based on multi-frequency ultrasonic technology is proposed.A power transformer oil parameter detection system is designed and manufactured to realize fast and accurate detection of multiple transformer oil parameters.First of all,the principle of multi-frequency ultrasonic testing technology for transformer oil parameters is studied.By analyzing the principle of ultrasonic testing,three common methods of ultrasonic testing are obtained.By studying the fluctuation and atteghnuation characteristics of multi-frequency ultrasound in transformer oil,the relationship between ultrasonic acoustic parameters and transformer oil parameters is obtained,and the technical principle of multi-frequency ultrasonic testing technology for transformer oil parameters is summarized.Secondly,based on the research of the multi-frequency ultrasonic detection method,a multi-frequency ultrasonic detection hardware system for transformer oil parameters is designed and fabricated.The measurement methods of ultrasonic velocity and attenuation coefficient are studied,and the design idea of hardware system is determined.The ultrasonic transducer and data acquisition card are selected,and the driving circuit and echo processing circuit of multi-frequency ultrasound are designed.Thus a hardware system of multi-frequency ultrasonic testing for transformer oil parameters is formed.This system can realize the transmission of different frequencies of ultrasound and the collection of echoes.It can acquire the parameters of ultrasonic speed and attenuation coefficient,and transmit the detection data to the upper computer software,thereby realizing the efficient and accurate detection of transformer oil parameters.Finally,a transformer oil parameter detection software system is designed and developed.A multi-frequency ultrasonic testing model for transformer oil parameters is established by using BP neural network which is optimized by improved grey wolf algorithm training samples.Based on this testing model,a software system for transformer oil parameter detection is designed and developed,which can process multi-frequency ultrasonic testing data,quickly generate the test results of transformer parameters,analyze the test results and give the operation status of transformer oil.The system realizes the efficient and accurate detection of the transformer oil parameters,which can play a good guiding role in the actual maintenance work.更多还原