| At present, the global most abundant clean energies include wind,solar, geothermal, and so on. Wind energy has been wholesale developed all over the world, the most effective conversion and use forms is wind power generation. The continuous development of large equipment manufacturing industry and wind power technology makes the wind turbine generator(wind turbine) keep developing in direction of large-scale, ocean, and the installation amount is increasing year by year.The total installed capacity of the world has reached 369,553 MW by 2014.The total installed capacity of the wind turbine is improving, and the occurrence rate of the faults and accidents is increasing, which leads to serious economic losses, and effect the grid security. In operation of the unit, there are some failures of the wind turbine which are blades, the gear box, the generator and the yaw system.As the input component of the wind turbine energy, the blade has serious influence on the operation and safety of the wind turbine. This paper analyzes several common faults and generating mechanism of the blades, and introduces several detection techniques of the wind turbine blades, and designs a set of monitoring system of the blade vibration signal.The hardware of the system includes strain gauge, amplifier, single chip,wireless transmission module and power supply module. Designed by LabVIEW software, the vibration signal of the blade can be monitored and displayed from upper computer system in real time.In view of the vibration signals of three kinds unmoral states and normal state, the signals of speed in 0.1rad/s is extracted by using the harmonic wavelet packet. An improved harmony search algorithm isproposed to optimize the SVM parameters, and the eigenvectors are trained and predicted by using basic HS-SVM and improved HS-SVM, the experimental results show that the improved HS-SVM has better classification results than the basic HS-SVM. |
PDF Full Text Request