模糊模式识别技术在超声表征铝质高压绝缘子中的应用

APPLICATION OF FUZZY PATTERN RECOGNITION TO HIGH VOLTAGE ALUMINOUS INSULATOR IN ULTRASONIC CHARACTERIZATION

本文提出一种基于模糊模式识别技术对不同烧成温度的高压铝质绝缘子进行超声表征的方法。通过超声水浸线聚焦技术获得烧成温度分别为1220℃、1260℃和1300℃的绝缘子的回波信号,并提取归一化功率谱中0.58MHz、1.282MHz和1.495MHz三个频率的幅值作为表征绝缘子组织结构的特征值。由于对应不同烧成温度,该特征值围绕其平均值波动,此时可认为该烧成温度对应组织结构的超声回波功率谱特征值形成了一个模糊集,利用模糊识别中能够表征模糊集之间彼此接近程度的贴近度概念来进行研究。结果表明,利用待测试样与标准试样功率谱特征值之间的贴近度,可以有效区分不同烧成温度的绝缘子,贴近度越大,待测试样与标准试样的烧成温度越接近。该方法为无损表征不同组织结构的绝缘子提供了有效途径。

Ultrasonic characterization method of high voltage aluminous insulator with various sintering temperature is developed based on fuzzy pattern recognition.The immersed line focus pulse-echo technique is applied to obtain echo signals of insulator with various sintering temperature.The power spectrum amplitude of characteristic frequency was chosen to evaluate the microstructure of insulator.For the insulator with sintering temperatures of 1220 ℃,1260 ℃ and 1300 ℃,the characteristic frequencies are 0.58MHz、1.282MHz and 1.495MHz.Corresponding to different temperatures,the amplitudes fluctuate around their average values,thus it is difficult to recognize the microstructure of insulator with various sintering temperature accurately.The amplitudes were considered as a fuzzy set,and closeness degree of fuzzy pattern recognition was used to evaluate the approaching degree among the fuzzy sets.Greater closeness degree corresponds to closer sintering temperature between the reference samples and the analysis samples.This method provides an effective approach for the evaluating of insulator microstructure.