新能源大规模并网、特高压交直流输电和智能电网的快速发展，使电网在源、网、荷各环节呈现电力电子化特征。新形势下电力信号具有宽频域、快变化、多时空尺度等新特性，传统测量方法难以满足监测要求。本项目拟从分数域多尺度分析角度研究电力电子化电力系统中宽带多频信号的高精度快速测量方法，主要研究内容如下：①基于深度置信网络和经典电力信号测量方法建立宽带多频信号数学模型。②基于多孔非下采样算法研究具备平移不变性的离散分数阶小波变换算法及其多尺度特性影响机理，辅以噪声自适应抑制算法研究，实现宽带多频信号的高精度测量。③基于高阶能量分布研究信号在分数阶小波变换子带内最佳分数阶变换阶次的快速求解方法，而后基于路由切换控制算法解决子带间最佳变换阶次的快速切换问题，实现宽带多频信号的快速跟踪。本项目将为电力电子化电力系统宽带多频信号的高精度快速测量提供理论支撑，对促进电网扰动准确监测及电网安全稳定运行具有重要意义。

With rapid development of integration of renewable energy, ultra high voltage ac/dc transmission and smart grid, the power system is increasingly characterized by power electronics in the "source-grid-load" segments. In the power electronized power systems, the power signals are characterized by wide frequency band, fast varying and multi-scale. The existing methods can't meet the measuring requirements. Based on the fractional multi-scale analysis, this project will research on the accurate and fast measurement of wideband and multi-frequency signal in power electronized power systems. Firstly, the mathematical model of the signal will be established by using the deep belief network and classical measurements of power signals. Secondly, based on the a'trous algorithm and nonsubsampled filter bank, a new form of discrete fractional wavelet transform will be proposed. And the factors which influence the multi-resolution performance of fractional wavelet transform will be studied. Then the adaptive noise suppression method will be studied to realize accurate measurement of the signal. Thirdly, fast detection method of the optimal transform order of fractional wavelet transform will be designed by studying the higher order energy distribution. Then the fast tracking of electrical parameters will be realized by using routing algorithm. This project will provide theoretical support for the accurate and fast measurement of wideband and multi-frequency signal in power electronized power systems, which is of great significance to the accurate monitoring of disturbances and the stability of power systems.