一种基于人工神经元的实时谐波及无功电流数字检测方法

A REAL-TIME APPROACH TO DIGITAL DETECTION OF HARMONICS AND REACTIVE CURRENTS BASED ON ARTIFICIAL NEURONS

在自适应干扰对消理论、人工神经网络理论的基础上,提出了应用人工神经元进行谐波、无功电流和谐波电流的自适应数字检测新方法。由于单神经元的自适应谐波及无功电流检测方法无论在结构上还是在权值迭代上都比较简单,为了实现实时检测,采用了基于神经元自适应谐波及无功电流检测方法。在实验中CPU为Pentium-100、内存为8M的PC兼容机用于数据计算,PCL-818L多功能卡作为集数据采样、A/D转换和数据传输硬件。实验结果表明所提方法是正确的,也证明了所提出的神经元自适应谐波及无功电流检测方法具有良好的实时性。

Based on the theories of adaptive interferencecancellation and artificial neural network, a new method foradaptive digital detection of harmonic, reactive current andharmonic current by use of artificial neurons is put forward.Because the adaptive detection of harmonic and reactivecurrent based on single neuron is rather simple either in the structure or in the iteration of weights. Therefore, to implement the real-time detection, an adaptive detection method forharmonic and reactive current based on artificial neuron isadopted. In the experiments a compatible PC withPentium-100MHz CPU and 8M Ram is used for calculation and a PCL-818L multi-function card is used for data sampling, A/D conversion and data transmission. The experimentalresults show that the presented method is effective and theobject of real-time adaptive detection of harmonic and reactive current are achieved.