Research On Data Analysis And Evaluating Models Of Collective Data From Hybrid Power Stations Of Wind Solar & Energy-Storage

The research is based on National 863 project "multi type reservoir can coordinate system control technology and demonstration," and provides an effective algorithm library aimed at data analysis of the wind, solar power station, which can not only processes with the output data of these stations effectively, but also gives quantitative assessment and research of the micro power system resolving the inherent difficulties among the wind, solar power stations, such as scene complexity, huge amount of data, fluctuation and diversity.VC++developed a set of data preprocessing and data analysis algorithm library, including: singular point detection, singular point repair, data normalization processing, clustering analysis, principal component analysis, wavelet algorithm, etc. In addition, an application program interface based on MFC has been developed.This article chose the measured data of Baoding Yingli PV demonstration power plant as a case for analysis. Based on study of the distribution law of the time scale and the amplitude of the photovoltaic power generation in the whole year, we get the conclusion:after clustering the similar days with clustering algorithm to 100 days’measures datas, It can separate the three kinds of weather:sunny, cloudy and rainy. Choosing the weather as the evaluation criterion, the accuracy rate is about 87.8%.Based on the data of similar days, the output fluctuation of cloudy days are Frequent and large amplitude from and the Short term output maximum fluctuation can be over 60% of the installed capacity. Then, we can get the change rate distribution histogram of 5 mins from probability density; we can see that, when the confidence interval is 99%, the maximum fluctuation rate is 12%. About 70% of the fluctuation degree of a whole year is less than 5%/min,12%/30min.After the analysis of the output characteristics of photovoltaic power station, the paper researched the energy storage allocation which was used in Tuoji Island intelligent island power station with the wavelet algorithm analysis. In this paper, two aspects of micro grid energy storage configuration were studied. Firstly, when the micro grid system was absorbed into the grid system, the grid connected component and the wave component were obtained to meet the requirements of the large power grid based on the domestic wind farm active power grid standards, according to the requirements of power fluctuation into large power grid to stabilize, by using the method of one order correction low-pass filter and grid connected power grid into wave correction. And the fluctuation component is obtained by using the method of wavelet decomposition to obtain the power signal of each storage, as a hybrid energy storage allocation basis. Secondly, when the micro grid system is in the isolated network system, the micro grid is a kind of typical isolated network system. In order to realize the load balance of power generation in micro grid, it needs to configure energy storage and diesel generator system for internal regulation with the use of these two indicators of electricity load rate and power supply reliability as constraint conditions. The power and capacity of diesel generator and energy storage system can be realized simultaneously with power allocation at all times in the configuration process and the increase of the time dimension gradient.