From Landscape to Portrait: A New Approach for Load Curve Data Analysis and Cleansing

Load curve data is critical in the demand-response management of smart grid. The quality of load curve data, however, is hard to guarantee since the data is subject to communication losses, meter malfunctions, and many other impacts. We present a new approach to organize the load curve data so that many tasks such as data visualization, outlier detection, and data cleansing become easy. Our method adopts a new view, termed portrait, on the load curve data by analyzing the periodical patterns in the data and re-organizing the data for ease of analysis. We introduce algorithms to build the virtual portrait of load curve data, and demonstrate its application on load curve data cleansing with real-world trace data. Compared to existing methods using regression-based time series analysis, our method is much faster and more accurate for both small-scale and large-scale datasets.