【作者】 郑常宝；

【导师】 王群京；

【作者基本信息】 合肥工业大学 ， 电力电子与电力传动， 2007， 博士

【摘要】 近年来，由于非线性负荷和感性负荷的大量使用，使得电力系统中无功功率不平衡越来越严重，导致电压偏差和电压波动等电能质量问题，直接影响电网的安全运行。静止无功补偿器(SVC)和新型无功发生器(ASVG)是电力系统中无功功率补偿的主要装置，而无功功率测量是实现无功功率补偿的关键技术，如何快速准确地获得无功功率一直是电气测量领域的研究热点。本文从Budeanu定义的无功功率出发，在较系统地研究了无功功率的测量方法的基础上，研制了静止无功补偿器模拟试验装置。主要工作和创新点如下：1．提出了正交小波变换和希尔伯特变换相结合的无功功率测量算法。该算法利用Hilbert变换没有移相误差和频带宽度限制的特性，将各次谐波电压分别平移90度。对电流和移相后的电压进行小波分解，从而在时间域求出各频段的无功功率。仿真实验表明该算法具有较高的测量精度。2．提出了小波包测量Budeanu定义无功功率的算法。利用小波包对信号的高、低频部分都具有多层次划分的性质，对电流和电压进行分解，由分解系数求出有功功率和无功功率等参数。仿真实验验证了该算法的有效性。3．提出了基于BP神经网络补偿TCR基波等效电纳与控制角非线性的方法。该方法在SVC控制系统的控制器和TCR之间插入BP神经网络，利用BP神经网络的非线性映射特性补偿TCR基波等效电纳与控制角的非线性。仿真结果表明该方法可以得到较好的补偿效果。4．提出了利用三角函数正交性测量基波无功功率的递推算法。根据三角函数的正交性和对称性求解基波参数，采用递推算法快速计算基波无功功率。该算法已在研制的静止无功补偿器模拟试验装置中得到应用验证。5．给出了一种基于小波变换的谐波滤除方法。该方法通过对负载电流进行小波分解，获得负载电流的低频分量，由低频分量重构得到基波电流，从而滤除负载电流中的谐波。通过对新型无功发生器仿真实验，表明了该方法能很好地滤除负载电流中的谐波。6．设计了静止无功补偿器模拟试验装置的主电路和以2407DSP为核心的控制电路的硬件和软件。对模拟试验装置进行了测试分析。无功发生变化时，模拟试验装置的最短调节时间为110ms。更多还原

【Abstract】 Recently, the use of many inductive and nonlinear loads in power system bring on more and more serious unbalance of reactive power that result in power quality problem, such as voltage deviation and voltage flicker, which affect the safe operation of electrical network. Static var compensator (SVC) and advanced static var generator (ASVG) are primary reactive power compensator, and the reactive power measurement is the key for reactive power compensator, therefore fast and accurate reactive power measurement is the focus in electric measurement.According to reactive power defined by Budeanu, this dissertation develops the simulative experiment equipment of SVC based on system research on reactive power measurement. The main work and innovation of the dissertation are as follows:1. A new algorithm is proposed using wavelet transform and Hilbert transform to calculate the reactive power. Hilbert transform can shift 90°phase for each frequency and has no error in theory. Current and voltage are decomposed by orthogonal wavelet transform, and reactive power in each frequency band can be calculated in time domain. Simulation result shows the highly precision of the algorithm.2. An algorithm of reactive power measurement using wavelet packet is proposed and simulated. Wavelet packet can decompose a signal into low frequency component and high frequency component, which can be further decomposed. Active power and reactive power can be calculated by the decomposition coefficient of current and voltage. Simulation proves the algorithm is useful.3. The method is proposed using BP Artificial neural network (ANN) to compensate nonlinear relation of the fundamental impedance of TCR with trigger angle. ANN is placed between controller and TCR in SVC control system, and the nonlinear relation is compensated by nonlinear map of ANN. Simulation result shows that the compensation effect is very well.4. An iterative algorithm based on an orthogonal characteristic of trigonometric function is proposed to measure fundamental reactive power. Fundamental parameters are calculated by symmetrical and orthogonal characteristic of trigonometric function, and the iterative algorithm can calculate fundamental reactive power fast. Simulation result shows that the precision of the algorithm is high. This algorithm is proved by the simulative experiment equipment of SVC.5. The dissertation presents a method of filtering out the harmonics based on wavelet transform. The low frequency component is obtained by wavelet decomposition of load current. The low frequency component is reconstructed to get fundamental current. Simulation result of ASVG shows the effect of filtering out the harmonics of load current by wavelet is excellent.6. The main circuit, the hardware and software of TMS320LF2407 DSP of the simulative experiment equipment of SVC are designed. The simulative experiment equipment is tested, and the shortest adjust time is 110ms when reactive power changes.更多还原