智能电网电力变压器噪声测量分析系统的设计与应用

发布时间：2019-05-28 17:07

【摘要】：近年来,随着人们对电能质量和电网安全的需求越来越高,智能电网的发展速度越来越快。智能变电站作为电能传输过程中的中转枢纽,发挥的重要的作用,而智能变电站中的主要设备电力变压器能否可靠运行关系着整个电网的安全。所以对电力变压器的状态进行监测显得尤为重要。变压器的运行状态与变压器的噪声密切相关。本文主要设计了一套噪声测量分析系统,对变压器的噪声进行采集,计算其声压级,观察分析变压器的噪声特性以达到实时监测变压器运行状态的目的。论文主要从噪声测量计算平台的设计、噪声信号的声级计算、噪声信号的分析三个方面对该系统进行设计。首先,对噪声测量分析平台进行整体设计,包括硬件的选型和软件任务流程的设计,其中硬件设计包括噪声数据采集模块、运算处理与控制模块、储存模块、通讯模块,软件架构主要包括FPGA控制数据采集流程,CPU计算任务的运行流程的设计。其次,基于IIR数字滤波器设计了A、C、Z频率计权网络,并通过MATLAB仿真得出其误差符合1级允差标准,适用于电力变压器噪声声压级的高精度计算。将实数快速傅里叶变换算法应用于噪声的频谱计算,相对于普通快速傅里叶算法,其计算量更少。通过观察变压器噪声信号的声压级与频谱值,可以实时监测变压器的运行状态。然后,在后台对噪声信号进行进一步的频谱分析,提出了用改进的1/3倍频程方法表示FFT,即用一段频带内的最大谱值表示该频带的谱值,用MATLAB仿真验证了其可行性。针对FFT的频谱泄漏现象,通过加窗函数减少误差,之后通过补偿添加窗函数引起的衰减值,提高了FFT分析的精度。针对FFT在频谱分析中的缺陷,提出了小波变换用于噪声信号的频谱分析,仿真分析了小波分析用于噪声信号频率成分的检测、噪声信号奇异性的检测。小波分析不仅能够进行频域成分的分析,还能分析信号的时域特性,为电力变压器噪声信号的故障分析提供了一种有效的分析手段。最后,利用本文设计的噪声测量分析平台采集得到现场的噪声信号,根据本文介绍的算法对实际信号进行计算分析,验证算法的实用性。证明本系统完全适用于变压器噪声特性的计算与分析。
[Abstract]:In recent years, with the increasing demand for power quality and power grid security, the development of smart grid is getting faster and faster. As a transit hub in the process of power transmission, intelligent substation plays an important role, and the reliable operation of the main equipment power transformer in intelligent substation is related to the security of the whole power grid. Therefore, it is particularly important to monitor the state of power transformer. The operation state of transformer is closely related to the noise of transformer. In this paper, a set of noise measurement and analysis system is designed to collect the noise of the transformer, calculate its sound pressure level, observe and analyze the noise characteristics of the transformer in order to monitor the running state of the transformer in real time. This paper mainly designs the system from three aspects: the design of noise measurement and calculation platform, the sound level calculation of noise signal and the analysis of noise signal. Firstly, the overall design of noise measurement and analysis platform is carried out, including the selection of hardware and the design of software task flow. The hardware design includes noise data acquisition module, operation processing and control module, storage module and communication module. The software architecture mainly includes FPGA control data acquisition flow and CPU computing task operation flow design. Secondly, the A, C, Z frequency weight network is designed based on IIR digital filter, and the error accords with the level 1 tolerance standard by MATLAB simulation, which is suitable for the high precision calculation of noise sound pressure level of power transformer. The real fast Fourier transform algorithm is applied to the spectrum calculation of noise, which is less than the ordinary fast Fourier algorithm. By observing the sound voltage level and spectrum value of transformer noise signal, the running state of transformer can be monitored in real time. Then, the further spectrum analysis of the noise signal is carried out in the background, and an improved 1 鈮，

本文编号：2487222