1 October 2017 in CIRED - Open Access Proceedings Journal
Here, a methodology based on cluster analysis, feature selection, and support vector machine (SVM) is presented to predict power quality indexes such as harmonic distortion and voltage deviation. First, cluster analysis is made on power quality index data, and historical data sets to be used are selected according to the classification of the data to be predicted. Then, with the feature selection algorithm, the optimal factor combination relevant to the power quality index to be predicted is obtained from several factors, such as weather, active power, reactive power, and others. Finally, the historical data and associated factor data are taken as the training sample data sets to establish the SVM prediction model. Through the analysis of the actual sample data during 2013 and 2014 in a certain area of Shanghai, China, it is proved that the average relative error of this method is greatly improved.
1 October 2019 in IOP Conference Series: Earth and Environmental Science
Echelon use batteries from electric vehicles will bring not only the cost reduction of energy storage but also the social benefits of circular using of resource, energy conservation and emission reduction. It is an important echelon use orientation that retired batteries from electric vehicles are rebuilt into distributed energy storage systems. The article introduces 8 cases of distributed energy storage systems containing echelon use batteries, whose application scenarios include load shifting, renewable energy storage, frequency modulation of power system, and capacity charge management. In summary, the reconstruction of echelon use batteries is based on battery packs or modules in order to reduce the cost of their secondary development as much as possible because the advantage of echelon use batteries is low cost after all. Echelon use batteries should be used under the regime of low rate and / or shallow charge-discharge for the sake of their cycle life and safety. Besides, a set of distributed energy storage system containing retired batteries from ROEWE e50 electric vehicles was developed by us and its application effects in a distributed PV generation station was introduced. Ten 3P3S retired batteries modules with similar capacities are rebuilt a 5 kWh energy storage system with 50 V rated voltage, whose energy transition efficiency is about 88%. Energy storage applications of retired batteries from electric vehicles in distributed PV generation not only supply electrical loads with stable power, but also achieve peak shaving of solar electric power.