Study On Wind Speed Characteristic And Uncertainty Model Of Wind Farm

At present, wind energy reserves is very rich and has become one of first choice for the development of new energy support ed in China. However, the effect of large scale wind power uncertainty becomes more obvious and the problem of abandon wind power rationing becomes more serious. Recent research on the uncertainty of wind power mainly concentrated on the statistical analysis of hourly average wind speed forecast and error pred ictions, while the long-circle characteristics study cannot meet the need of real-time scheduling and controlling in the power grid. In this passage, wind speed characteristic and uncertainty model are studied at the view of power grid’s real-time scheduling and optimization controlation, which can provide a theoretical foundation and supporting evidence for large-scale wind power.Firstly, according to the existing power grid real-time scheduling and optimization control relationship,and the atmosphere movement general gap time scale,decompose and compound actual wind data.Then found the multiscale modulation effect of wind amplitude-frequency,and got the instantaneous spectral characteristic.After analyzing corresponding physical mechanism,we found that the not only wind speed has the multi-scale amplitude modulation effect,the hourly mean wind speed also has the modulation effect in the turbulent wind speed frequency. We developed the three parameter power law model of the turbulence intensity, which further improves the IEC standard. According to the amplitude modulation effect of hourly wind power on residual wave component and instantaneous power spectrum of wind power, the time-frequency separation characteristics of wind power was obtained. According to analysis of wind’s turbulence nature, the corresponding physical mechanism is analyzed.Secondly, it’s found that there is daily modulation effect in the wind power variance, and the change pattern of diurnal cycle is studied. The mean turbulence intensity corresponding to the effective operation of the wind speed for wind turbine acts the average turbulence intensity corresponding to the height of the wind turbine hub. On the basis of considering the factors such as season, region and height, the diurnal cycle of wind power variance is studied. A time varying modeling method is proposed, which can eliminate the influence of diurnal cycle and improve the fitting accuracy of the model.Similarly, the wind power also has a daily cycle variance. Then, in view of the physical mechanism of the diurnal cycle and its practical significance for the safe and efficient use of wind power, the necessary analysis and case study is carried out. An example of frequency modulation capability analysis is given to the above applications by using twoarea and four-machine power system. The simulation results show that: considering the time variant model of diurnal cycle in wind power fluctuation, the model can meet the scheduling staff to configure the capacity of FM unit well, and satisfy the requirement of the actual frequency modulation capability.Thirdly, also with one hour time scales, modulation effect in wind speed and wind power variation was found using variation analysis to study on the wind speed and wind power. Further, a three parameter model that can quantitatively describe change rate in wind speed and wind power. Then, a modeling method system concerning about wind power uncertainty was formed corresponding with chapter two. Meantime, it’s found that the variation of wind power also has the characteristics of diurnal cycle. At last, the strategy of integration stabilization is proposed, which could considered its cost by tracking the different frequency of wind power forecasting curve error component based on the integrated power generation unit to realize the planning ability of power production and reduce spinning reserve. And the accommodating ability of wind power and power grid security can also be further improved.Fourthly, a quantitative description method based on correlation function is proposed for the predictability, which is one kind of characteristic for wind speed uncertainty. On the basis of the above, the scale of the information effectiveness that can used for wind speed forecasting is quantitatively analyzed and the optimization framework of the forecast of average wind speed is presented. Considering the important role wind speed forecast plays in the use of large scale wind power integration, the predictability of wind speed variation and variance was also studied and the predictable length of them is obtained. Then cases study is analyzed by using support vector regression and deep learning algorithms. Thus, a new multi-parameter prediction frame of wind speed is proposed under a new understanding of predictability.