GP法在发电机转子绕组匝间短路诊断中的应用

Genetic programming for diagnosing turn-to-turn short fault of generator rotor windings

用遗传规划法建立发电机转子励磁电流与机端电压、有功、无功的数学模型。发电机在转子绕组发生匝间短路后,励磁电流会增加,为了维持磁动势平衡,对应于某一确定的运行工况,发电机的励磁磁动势会维持不变,发电机的励磁电流和机端量之间的关系可以反映转子绕组的状态。以机端量为自变量,遗传规划能自动生成函数表达式来体现它们之间的关系。通过实例计算并与神经网络方法相比较,验证了该方法的有效性。

A new method based on genetic programming (GP) for simulating the relationship of the field current of generator with the voltage, active current and reactive current is presented. EMF will be constant on a fixed condition whereas the exciting current increases to maintain the balance of magnetic potential when the rotor winding shorted. Then the relationship of field current with voltage, active current and reactive current can reflect the rotor state. Using the terminal variables as independent variables, GP can automatically create a function to reflect their relation. The result of a computation example shows that the method is effective.