摘要： 针对船用污水处理装置运行状态的监测问题，提出一种含时间因子的正交状态基设备运行状态分类模型。根据设备系统的运行特点、条件及各状态的时间属性，确定训练集变量的取值特征，将输入向量映射到一个正交完备的特征空间，生成各数据类别的正交状态基，由此得到判别矩阵，并引入时间相位因子构成含时正交状态基矢量，从而实现对未知状态的分类。对船用污水处理机13种运行状态的实测样本进行检验，实验结果表明，该模型能有效监测处理机的运行过程，正确区分不同的故障模式。

关键词: 正交状态基, 设备运行状态, 污水处理, 时间相位因子, 故障模式

Abstract: A new model to design data classifiers by using orthogonal state basis with time is presented. The space of input data is extended to a relevant complete orthogonal feature subspace according to the unique characteristics of the input variables, results in the formation of categorical orthogonal state basis and the classification matrix. The unknown input data is projected onto the subspace through the matrix to be sorted, the time variable is introduced as an exponential factor and the duration of a state lasted is calculated, and the second classification based on the time duration of a state is performed. It applies experimentally such model to a boat sewage purifier for the classification of its 13 working states, and the results indicate that the fault modes can be effectively discerned and the behaviors of the machine system can be monitored and properly treated.

Key words: orthogonal state basis, equipment operational condition, sewage treatment, time phase factor, fault mode

中图分类号: 

• TP181