一种改进的模糊控制方法及其在大坝安全上的应用

大坝的安全运行状态影响因素多变且各因素之间关系错综复杂 ,而目前大坝普遍存在事故隐患以及对其安全所使用的监控方法的落后和效果的不良 ,笔者针对上述原因 ,给出了一种新的应用模糊理论的算法 ,将传统的模糊控制理论用神经网络优化后 ,能更加方便地应用于非线性系统建模 ,利用给出的新的学习规则 ,可以用竞争学习为训练样本的输入空间进行聚类 ,然后为其确定区域划分边界。将该法运用于水坝的安全管理上 ,将得到的训练网络用实际的历史记录数据验证。结果表明差距很小 ,证明了该方法的正确性 ,将其应用于大坝的安全监控和事故预测方面将会有很好的发展前景。

A Modified Fuzzy Control Method and Its Use in Dam Safety

In order to secure the dam many factors are involved,and the correlation among these factors are very complicated. At present the hidden dangers widely exist and their control measures are rather ineffective. A new algorithm for learning fuzzy rules is presented. After optimized by the artificial neural network, the fuzzy theory could be used for the nonlinear system modeling more easily. The algorithm first partitions input data into some clusters by competitive learning, and then the decision margins for each input cluster are determined. And finally, the fuzzy rules for each input local region are learned. The new learning rules are used to secure the dam. The observed historical data of the dam is provided to verify the new rules. The result shows that the prediction is valuable with good perspective.