电力系统谐波分析方法研究

发布时间：2019-03-08 20:48

【摘要】：由于大规模非线性元件以及设备再电力系统的广泛应用,导致大量的谐波注入电网,严重影响了电网的电能质量,因此谐波治是非常必要的。但无论是谐波责任认定还是谐波的治理,都依赖于谐波的特性。快速傅里叶变换(FFT)是谐波分析常用的方法。但电力系统的谐波并非精确50Hz,且存在大量的间谐波,故导致信号非同步采样,并且由于栅栏效应、混叠效应、截断效应,导致无法精确的估计出信号的幅值、频率和相位。另一方面,FFT的每个谱峰都不可避免的受到泄露的影响,故定量的给出每个谱峰的可信度也是一个关键问题。考虑到谐波是时变或者瞬时性的信号,无法进行较长时段的采样,故直接使用FFT,导致信号分析的频率分辨率低。本文针对这三个问题展开,详细的阐述了在减小计算量的同时,如何有效的提高参数估计的精度和可靠性。研究了谱峰可信度的一些性质。针对传统的谱峰可信度判断方法的缺陷,本文借助一族偏差公式给出商集,通过比较这些商与1的接近程度,进而给出更精确的谱峰可信度;提出了一种基于状态空间模型的谐波参数估计方法。首先建立了谐波信号的状态空间模型,为了提高参数估计对噪声的鲁棒性,通过分析将参数估计的鲁棒性问题转化成一个求酉矩阵的过程,最后通过求该酉矩阵特征值即可高精度的估计出谐波频率值。幅值和相位则可以通过最小二乘法或者其他优化方法得到。另外还研究了状态空间法和Prony法的联系以及有效奇异值选取问题;针对同时含有谐波和间谐波的谐波分析问题,采用基于FFT和状态空间法的谐波分析方法。为了提高算法的效率和可靠性,先通过判断FFT的谱峰的谱间干扰程度分离出能用FFT分析的频谱分量,进而借助能量重心法估计出对应的分量的参数;接着利用状态空间法估计出剩下谐波参数。在此基础上又提出了两种降维技术以提高算法的实用性。
[Abstract]:Due to the extensive application of large-scale non-linear components and equipment re-power system, a large number of harmonics are injected into the power grid, which seriously affects the power quality of the power grid. Therefore, harmonic treatment is very necessary. However, whether it is the recognition of harmonic responsibility or the treatment of harmonic, it depends on the characteristics of harmonic. Fast Fourier transform (FFT) is a common method for harmonic analysis. But the harmonic of power system is not accurate 50Hz, and there are a lot of inter-harmonics, so the signal is not synchronous sampling, and because of fence effect, aliasing effect and truncation effect, it is impossible to accurately estimate the amplitude, frequency and phase of the signal. On the other hand, each peak of FFT is inevitably influenced by leakage, so it is a key problem to give the credibility of each peak quantitatively. Considering that harmonics are time-varying or transient signals can not be sampled for a longer period of time. Therefore the direct use of FFT, leads to low frequency resolution of signal analysis. In view of these three problems, this paper expounds in detail how to improve the accuracy and reliability of parameter estimation while reducing the computational complexity. Some properties of the reliability of spectral peaks are studied. Aiming at the defect of traditional spectral peak reliability judgment method, this paper gives the quotient set by a family of deviation formulas. By comparing these quotient with 1, the more accurate spectral peak reliability is given. A harmonic parameter estimation method based on state space model is proposed. In order to improve the robustness of parameter estimation to noise, the robustness of parameter estimation is transformed into a unitary matrix. Finally, the harmonic frequency can be estimated with high precision by calculating the eigenvalues of the unitary matrix. Amplitude and phase can be obtained by least square method or other optimization methods. In addition, the relationship between the state space method and the Prony method and the selection of the effective singular value are also studied, and the harmonic analysis method based on FFT and the state space method is adopted to solve the harmonic analysis problem which contains both harmonics and inter-harmonics. In order to improve the efficiency and reliability of the algorithm, the spectral components which can be analyzed by FFT are separated by judging the interspectral interference of the spectral peaks of FFT, and then the parameters of the corresponding components are estimated by means of the energy center of gravity method. Then the state space method is used to estimate the remaining harmonic parameters. On this basis, two dimension reduction techniques are proposed to improve the practicability of the algorithm.
【学位授予单位】：湖北工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM711

【参考文献】