灵活的间歇生产是决定企业生存的重要因素，其在线监测、故障诊断及控制具有重要意义。项目以具有多时段、非高斯性、强非线性等特点的间歇过程为对象，重点开展统一框架下的非线性系统故障诊断与控制研究。与传统多变量统计监控的被动诊断不同，提出间歇过程的主动故障诊断思想，采用多模型结构设计合理辅助信号，充分激励隐藏的故障信息，提升检测精度；在此基础上构建主动故障诊断与控制器设计的统一架构，实现故障检测器和控制器的同步设计；从过程、质量变量同时出发，构建多方法融合的统计监控模型，确保对系统非高斯、非线性特性的提取，实现被动模式下的故障有效检测；在主、被动检测并行模式下构建微小故障检测双保险：以随机共振放大为指标设计主动诊断的测试信号，并利用结构信息构建残差发生器，区分噪声与微小故障信号。项目提出一套适合复杂非线性系统的主、被动故障诊断与监控的理论体系，其实施有助于提高我国间歇过程的生产力。

Flexible batch production is an important factor in determining the survival of enterprises, and the process monitoring of batch process is becoming urgent and necessary. This proposal focuses on the research of active fault diagnosis and control of batch process in an unified formulation. The active fault diagnosis of batch process, different from the traditional passive diagnosis multivariate statistical monitoring, is proposed, and a multimodel structure is used to design the reasonable auxiliary signal for fully stimulating the hidden fault information.Then an unified structure of fault diagnosis and controller design is built based on the active fault detection, and the fault detector and controller are synchronously desig; For the detection of obvious fault, fusion statistical method is given to construct monitoring model and extract infromation with Gaussian, non-Gaussian and nonlinear characteristics. Here process variables and quality variables are considered concurrently, which can be used to detect faults relative to production quality. Finally the parallel of active and passtive detection is adopted for incipient fault monitoring, in wihic the analytical models is fused with process data to discriminate noise from process data in passive detection mode,and the stochastic resonance criterion for auxiliary signal design is used to amplify the incipient fault in active detection mode. Implementation of all the proposed method will contribute in the statistical monitoring theory of batch process, and advance the automation level of domestic batch industry.