Gu Cailian

Evaluation of Abandoned Wind Power by Neural Network Method

Wind farm operation is characterized by large scale, fast and disorderly, the construction of grid structure is not complete, the fast adjustment of power supply in the power grid does not match to others, the objective laws of electric power system and uncontrollable and intermittent of wind power, and adjustable power supply capacity constraints in power grid, all of them cause the consumptive ability of wind power grid, which leads to more and more abandoned wind. This paper studies the tower neural network method, according to different height wind speed and wind direction data of historical wind tower measuring, combining with the fan power of historical observation data of wind farm, and building the neural network model, then the sample data will be input to the neural network model which has built to get theoretical power fan and abandoned wind power. By comparing the wind tower method, neural network method, model machine method with area integral method calculate abandoned wind power data, evaluate the effect that based on abandoned wind power tower evaluation model of neural network method in low wind speed has a good reference value, relatively close to the measured wind speed.

The sensorless control technology based on rotor initial position detection

This paper studies on sensorless control technology applied in permanent magnet synchronous motor. With respect to the problems in existing test method for initial position, the paper starts with the technology to observe flux linkage and detect initial position used in sensorless control technology in direct-drive permanent magnet wind driven generator, and on this base, a theory of detecting rotor initial position by mixing two-phase with three-phase for breakover is put forward. After simulation analysis, the correctness is verified, then on-site experiments are conducted on direct-drive permanent magnet wind driven generator twin trawling experimental platform provided by some cooperative institutions, with the direct-drive permanent magnet wind driven generator set on a condition with slow-speed of revolution. It turns out that the method of mixing two-phase with three-phase for breakover could not onlydetect the rotor initial position quickly and effectively, but also possibly control the error within ±15° electrical angle. At the same time, it is applicable for the low cost, easy to implement, which is capable of meeting the requirements to start a direct-drive permanent magnet wind driven generator.