基于参数区间的非线性动态系统故障隔离与辨识方法的研究

项目名称： 基于参数区间的非线性动态系统故障隔离与辨识方法的研究

项目编号： No.60864003

项目类型： 地区科学基金项目

立项/批准年度： 2009

项目学科： 电工技术

项目作者： 李泽滔

作者单位： 贵州大学

项目金额： 27万元

中文摘要： 对于一般性的非线性动态系统故障隔离和辩识，目前还缺乏有效的快速方法。建立在模型单调性特性假设基础上的基于参数区间的动态系统故障隔离和辨识方法是一种新方法，能适用于较多类型的非线性动态系统，且隔离速度快，并在完成故障隔离后同时也得到了故障参数的估计值，因此具有很好的发展前景。然而作为一种新方法，尚有一些问题没有解决，比如多重参数的故障隔离问题，模型参数不确定性问题，噪音问题等。该项目的目的就是研究这些问题。研究的关键是弱化甚至消除方法对模型单调性特性假设的依赖。项目的研究取得了预期成果。利用区间计算理论（Interval Arithmetic）实现了多重参数区间的系统模型响应输出包络线（Envelope）的计算，并以此建立了故障隔离与参数辨识机制。在新机制中模型单调性特性假设不再需要。新方法不但能够实现多重参数故障的隔离，还能应用于非单调性系统，模型参数不确定性问题也因此有了解决办法。项目还采用等张（Isotonic）理论对问题进行了研究并取得了初步成果，得到的多重区间判断曲线更紧凑。项目还对参数区间方法的许多相关问题进行了研究，比如自适应区间、传感器故障、执行器故障、容错控制等。

中文关键词： 故障隔离与辨识；参数区间；多重参数故障

英文摘要： For general nonlinear dynamic systems one lacks effective fast method for fault isolation and identification. Parameter intervals based method which based on the assumption of monotonicity property of model can be applied to many types of nonlinear dynamic systems and its isolation is fast therefore has good prospects. However, there are still problems should be studied, such as multi-parameters fault isolation, uncertainties of model parameters, etc. The objective of this project is to study these issues. The critical point of the project is to weaken or even eliminate the hypothetical dependence of model monotonicity property. The project has achieved the expected results. In the study, Interval Arithmetic theory is studied and is used to calculate the envelope curves of the system output curves family with respect to a multi-parameters interval, and this envelope is used to build the fault isolation mechanism. The monotonicity property is no longer needed. The new method can realize multi-parameters fault isolation, and can be applied to non-monotonic systems, the problem of model parameter uncertainties consequently can be easily solved. The study of the project also uses isotonic theory to solve the calculation of the judgment corves of multi-parameters interval and a preliminary result has been achieved, the obtained judgment corves are more compact. The project also carried out researches for many related issues to the parameter interval based fault isolation, such as adaptive interval, sensor fault, actuator fault, fault tolerant control, etc. The results of the project have great significance in theory and have good application prospects.

英文关键词： fault Isolation and Identification; Parameter Interval; Multu-Parameters Fault