基于神经网络与光纤传感阵列的结构状态监测方法

STRUCTURAL STATE DETECTION WITH FIBEROPTIC SENSING ARRAY AND NEURAL NETWORK SIGNAL PROCESSING TECHNIQUES 1)

以探测机敏复合材料与结构内应变、应力及损伤等物理状态的位置信息为目的．提出了一种新颖的可内埋于材料与结构内作为结构状态监测用的强度调制型光纤传感阵列网络，并采用人工神经网络来处理传感阵列输出的并行分布式传感信号，阐述了适用的Ｋｏｈｏｎｅｎ网模型及其变化形式。

In this paper, a type of simple and excellent crank type etched fiberoptic intensity sensor is presented. A three dimension sensing network, which is composed of etched fiberoptic sensors, is embedded in the laminated composite material structure of vertical tail wing of training 11 plane for the purpose of nondestructive evaluation of structural states in smart composite material and structures. The composite specimens in which sensing network is embedded are tested with stretch and bending in three points. The experimental results have shown the feasibility of embedded fiberoptic sensor network to be used to measure the structural states such as strain, stress and damage. It is also shown in experiment that the changes of optical power are highly dependent on the strain parallel to optical fiber axis and independent on the strain vertical to optical fiber axis; the responses of fiberoptic sensor are linear, repeatable, have a high sensitivity and no measurable hysteresis; the break threshold of etched optical fiber can be controlled by changing the etched time, and it is useful to promote the sensing sensitivity. For above fiberoptic smart structures where sensing and actuation is distributed over large areas or consists ofdozens to thousands of discrete elements, the processing task is computationally intensive. Artificial neural networks offer an opportunity to implement a massively parallel architecture with near real time processing speed and the ability to learn the desired response. In this paper, the Kohonen network and its changes which are suitable for distributing sensing signal processing in smart structures are described. The simulation results have shown that their ability for identifying the position of stress, strain or damage is above 90%.