基于人工神经网络测量含谐波无功功率的快速算法设计与实现-计算机技术专业论文.docxVIP

摘要 摘要 万方数据 万方数据 万方数据 万方数据 摘要 无功功率的测量对无功补偿、节能减排都有着重要的意义。但是近年来，随 着电力系统的发展，越来越多的现场存在谐波现象，如何准确测量无功功率尤为 重要。目前大部分智能电量仪表均采用以 ADE7758 芯片为核心的解决方案进行交 流采样，该芯片因功能全面、编程方便、硬件简单、价格便宜，深受广大用户喜 爱。但是由于 ADE7758 内部算法的固有缺陷，使其在测量含有谐波的现场环境中， 所计算出的无功功率存在明显偏差。论文首先分析研究了国内外学者提出的一些 算法，如希尔伯特算法、FFT 算法、小波变化以及处于研究阶段的人工神经网络 算法等。在此基础上，着重研究了神经网络在测量无功功率上的应用，进而设计 并实现了一种基于神经网络测量无功功率的快速算法。论文从网络模型选取、网 络结构设计、学习算法设计等多方面进行了神经网络设计与优化，提高了速度和 精度。论文对该算法进行了编程实现，通过编程仿真验证了算法的精准度和实时 性，并与其他的常用算法进行了对比分析。结果表明，该算法实时性好、精准度 高，很好地达到了预期的设计目标。目前，该算法已经投入到了新一代产品的研 发设计中。 关键字：谐波 无功功率 ADE7758 人工神经网络 Ab Abstract 万方数据 万方数据 万方数据 万方数据 Abstract Measurement of reactive power is very important for reactive power compensation and energy saving. In recent years with the development of power system, more and more harmonics phenomenon appears. How to accurately measure the reactive power is particularly important. Most of the current intelligent meters use an ADE7758 chip as the core solution to realize AC sampling. The chip delights most users by comprehensive functions, easy programming, simple hardware, low price. But the chip ADE7758 has obvious errors in the measurement of reactive power in harmonic field for its algorithm defects. This paper analysis some algorithms proposed by the domestic and foreign scholars, such as the Hilbert algorithm, FFT algorithm, wavelet transform and the artificial neural network which remains in research stage. Then the paper focuses on the research of the neural network application on measurement of reactive power, and has designed a fast algorithm successfully. In order to achieve a high speed and accuracy, the fast algorithm is optimized from network model selection, structure design, learning algorithm design and other aspects. The algorithm is realized by programming. The precision and real-time of the algorithm are verified though simulation programming, and then the paper compares with other commonly used algorithms. The result shows that this algorithm has a higher real-time and accuracy, and ac