随机分布序列互联系统的协作容错控制研究

The complex industrial process system always requires high reliability and safety, the models of these large-scale systems are mainly assumed to have non-Gaussian random disturbance and interconnected subsystems. The traditional fault tolerant control for single system cannot satisfy the security requirements of the interconnected system. Based on the above requirements, we propose an active fault diagnosis and.collaborative fault tolerant control methods regarding for an irreversible non-Gaussian stochastic distribution interconnected system with boundary conditions. Specifically, the main contributions of this project are listed as follows: 1) An active fault diagnosis method will be developed for small fault; 2) A distributed collaborative fault tolerant controller will be designed for interconnected stochastic distribution control system based on the constraint of communication cost; 3) A hierarchical collaborative fault tolerant controller based on optimal allocation strategy will be derived; 4) The collaborative fault tolerant control method will be applied to the control of density distribution of paper fiber in mechanical pulping process both in theory and in practice. The results in this project will provide novel directions for the research domain of collaborative fault tolerant control for interconnected stochastic distribution control system. And, in real world application, our results will provide a new tool for the safety and reliability of actual problem.

复杂工业过程具有系统规模大、对可靠性与安全性要求高、随机干扰多为非高斯分布且具有多子系统互联结构的特点。当系统发生故障时，传统的单一系统容错控制方法难以实现系统对安全性的要求，亟需研发针对随机变量为非高斯分布且多子系统互联情况下的故障诊断与容错控制方法。鉴于此，本项目针对一类受边界条件约束的不可逆非高斯随机分布互联控制系统，将研究其主动故障诊断与协作容错控制问题。具体研究内容为：1)针对微小故障或慢时变故障设计快速主动故障诊断算法；2)设计基于通信成本约束的随机分布序列互联控制系统的协作容错控制器；3)设计基于最优分配策略的分级协作容错控制器；4)基于非高斯随机分布序列互联系统的协作容错控制方法，实现其在造纸机械制浆控制过程中的仿真验证及实际应用。本项目在理论方面能够为随机分布序列互联系统的协作容错控制研究提供新的思路，在应用方面能够为实际问题的安全性与可靠性提供新工具。