CAJ | 학술논문

提出一种基于总体最小二乘法-旋转不变技术信号参数估计(TLS-ESPRIT)算法的间谐波参数估计方法.首先通过TLS-ESPRIT算法准确估计电网信号中谐波和间谐波的个数及频率,然后应用自适应神经网络来估计其幅值和相位.自适应神经网络用于谐波分析时,不需要事先进行训练,具有实现简单快速的优点.仿真结果表明该算法无需数据同步采样、频率分辨率高、抗干扰性强、谐波和间谐波参数估计准确.
제출일충기우총체최소이승법-선전불변기술신호삼수고계(TLS-ESPRIT)산법적간해파삼수고계방법.수선통과TLS-ESPRIT산법준학고계전망신호중해파화간해파적개수급빈솔,연후응용자괄응신경망락래고계기폭치화상위.자괄응신경망락용우해파분석시,불수요사선진행훈련,구유실현간단쾌속적우점.방진결과표명해산법무수수거동보채양、빈솔분변솔고、항간우성강、해파화간해파삼수고계준학.
An interharmonic parameter estimation method based on Total Least Squares Estimation of Signal Parameters via Rotational Invariance Techniques algorithm was proposed in this paper. The numbers and frequencies of harmonics and interharmonics in power network signals are estimated by TLS-ESPRIT algorithm, and then their am-plitudes and phases are analyzed by applying adaline neural network. ESPRIT algorithm was proposed by Roy et al in 1986, which can be applied to a wide variety of problems including accurate detection and estimation of sinusoids and direction of arrival of noise. It exploits an underlying rotational invariance among signal subspaee induced by an array of sensors with a translational invariance structure. The technique manifests significant performance and computational advantages over MUSIC algorithm. When Adaline neural network is used in harmonic analysis it need not train in advance and has the characteristics of simplicity and celerity. The simulation results showed the algorithm which needs no synchronous sampling data has high frequency resolution and strong anti-jamming capability, and the parameters estimation of harmonics and interharmonics is accurate.