Development of diagnosis technique using trained model

在这项研究中，我们开发了基于模型的诊断技术的旋转轴，发动机悬置和空调。首先，我们使用Modelica语言建立模型。然后，我们在一个模型中建立各种异常。用物理关系对正常现象和异常现象进行建模。其次，进行了有故障和无故障的模拟，以开发诊断技术。我们使用语义网络，逆技术和支持向量机的诊断。最后，我们证实我们开发的技术可以很好地区分异常病例和正常病例。

In this study, we developed the model based diagnosis technique for a rotational shaft, an engine mount and an air-conditioner. First, we build models by using Modelica language. And then, we build various abnormalities in a model. Modeling was carried out by Physical relations both for normal and abnormal phenomenon. Second, simulations were done with and without failures to develop diagnosis technique. We used semantic network, inverse technique and support vector machine for diagnosis. Finally, we confirmed our developed technique well distinguished an abnormal case from normal one.