【作者】 李敏；

【导师】 张迎春； 刘明；

【作者基本信息】 哈尔滨工业大学 ， 控制科学与工程， 2020， 博士

【摘要】 由于元器件自身的使用寿命和系统自身固有的特性等因素,系统中长期运行的元器件,例如执行机构和传感器,往往会发生一些未知的故障,这些故障可能会造成系统性能下降甚至导致系统是不稳定的。因此,在控制系统设计的时候,需要考虑控制器的容错能力,确保当系统在发生故障时,系统仍然是稳定的且具有满意的性能。容错控制技术作为一种有效的解决该问题的方法,受到许多人的关注。另一方面,由于系统中存在未建模参数误差和外部扰动,控制系统的鲁棒性也是一个重要考虑方面。滑模控制技术是一种能够有效消除模型参数不确定和外部干扰影响的鲁棒控制方法。因此,结合滑模控制技术和容错控制技术设计具有强鲁棒性的滑模容错控制算法具有重要意义。此外,为了降低控制系统相应模块之间的通信频率,在控制系统设计中引入事件触发通信机制。但是,事件触发机制的引入带来了一些新的研究问题,例如设计有效可行的事件触发规则保证奇诺行为不存在,同时设计滑模容错控制器克服事件触发机制造成的状态误差的影响,保证滑模面的可达性问题。基于上述所述,本文围绕相应的事件触发滑模容错控制方法开展深入研究,其具体研究内容包括:研究了一类执行机构故障的线性系统自适应状态反馈事件触发滑模容错控制问题。采用动态均匀量化器对传感器到控制器端和控制器到执行机构端的信号进行量化。基于量化的状态向量信号设计可行的事件触发条件。结合滑模控制技术和自适应控制技术设计了自适应状态反馈事件触发滑模容错控制算法保证闭环系统的稳定性和滑模面的可达性。给出了可以避免奇诺行为的事件触发条件参数设计方法,分析了任意两个连续触发事件之间的时间间隔。研究了一类执行机构故障的线性系统自适应输出反馈事件触发滑模容错控制问题。利用动态均匀量化器对传感器到控制器端和控制器到执行机构端的信号进行量化。设计一个动态补偿器,结合量化的测量输出向量和动态补偿器的状态向量设计了可行的事件触发机制。基于测量输出和动态补偿器状态向量设计了一种滑模面,并给出了滑模面稳定的充分条件。结合自适应控制技术和滑模控制技术,设计了自适应输出反馈事件触发滑模容错控制律保证滑模面的可达性。通过理论分析给出了事件触发条件参数的选择方法,得到了任意两个连续触发事件之间的时间间隔估计值。研究了一类执行机构故障的Markov系统异步自适应事件触发滑模容错控制问题。利用动态对数量化器对传感器到控制器端和控制器到执行机构端的信号进行量化。基于量化的状态信息,设计一种新的事件触发规则。结合滑模控制技术和自适应控制技术,提出一种异步自适应事件触发滑模容错控制策略,使其能够有效补偿执行机构故障、事件触发机制导致的状态延迟和量化误差对系统的影响,且同时能够保证闭环系统的状态轨迹到达滑模面附近的一个小邻域内并停留在邻域内。给出了所设计的事件触发条件能够避免奇诺行为的充分条件,并从理论上分析在所设计的事件触发机制下,任意两个连续触发事件之间的时间间隔。研究了一类执行机构故障和状态依赖的攻击信号情况下的Markov系统自适应事件触发滑模容错控制问题。对原系统进行坐标转换,利用新坐标下的状态向量设计滑模面,分析了滑动模态的稳定性,给出了滑模面参数的设计方法。基于新的坐标下的状态向量设计了可行的事件触发条件,给出了事件触发条件参数的选择方法。利用受损的事件触发状态向量和滑模面设计了自适应事件触发滑模容错控制策略,实现对未知参数的估计和补偿,保证系统的稳定性和滑模面的可达性,克服执行机构故障、事件触发状态误差和攻击信号的影响。分析了任意两个连续触发事件之间的时间间隔。研究了一类混杂执行机构故障和攻击信号情况下Markov系统自适应事件触发滑模控制问题。设计了一种模态依赖的滑模面,使得布朗随机过程对滑模运动的没有影响,给出了滑动模态的稳定条件。设计了一种可行的事件触发机制,并提出了一种模态依赖的自适应事件触发滑模容错控制策略,能够有效补偿执行机构故障、外部干扰、事件触发误差和攻击信号的影响,保证系统的稳定性和滑模面的可达性。分析了所设计的事件触发机制的最小触发事件间隔。更多还原

【Abstract】 Due to the service life of components,the inherent characteristics of the system and so on,some unknown faults,such as actuators failure,sensors faults and system faults,may often occur in control system,which will lead to that the system performance degrades and even the system is unstable.Therefore,it is required that the control system has fault tolerance ability to against these faults with maintaining an acceptable system performance.Fault Tolerant Control(FTC)has been shown as an effective approach to deal with this problem and many researchers’ interesting have been attracted.On the other hand,due to the existing parameter uncertainties and external disturbance,the robustness should be considered in control system design.Sliding Mode Control(SMC)is a powerful tool to eliminate the effects of model uncertain parameters and external disturbance.Hence,it is significant and important to develop fault tolerant and robust control strategies via the approach of combining sliding mode control and fault tolerant control techniques.Moreover,to reduce the communicate rate of system devices,event-triggered mechanism is introduced.However,the introduction of event-triggered mechanism also causes some problems needing to be further studied,for example how do we design an suitable eventtriggered condition to avoid the Zeno behavior,and how do we design a event-triggered sliding mode fault tolerant controller to overcome the effects of the state error induced by event-triggered mechanism and to ensure the reachability of sliding surface.Thus,in this work,the event-triggered Sliding Mode Fault Tolerant Control(SMFTC)design approach has been investigated.The detail contents of this work are listed as follows:The problem of adaptive event triggered SMFTC for a class of uncertain linear system with actuator failure and signal quantization via state feedback is studied.The classical dynamic uniform quantization strategy is employed to quantize data in both sensorcontroller side and controller actuator side.A new adaptive event-triggered SMFTC scheme via state feedback is proposed to stabilize fault closed-loop systems,where the event-triggered condition is based on quantized state vectors.The reachability of the proposed sliding surface can be ensured by the designed control scheme.Furthermore,sufficient conditions under which zeno behavior can be avoided is given,and the interexecution time interval is presented.The problem of adaptive event-triggered SMFTC for a class of uncertain linear system with actuator failure and signal quantization via output feedback is investigated.The classical dynamic uniform quantization strategy is employed for data quantization in both sensor-controller side and controller-actuator side.An event-triggered condition is established based on quantized augmented state vector that includes output measurement vector and dynamic compensator state vector.A new adaptive event-triggered SMFTC scheme via output feedback is proposed to stabilize the fault closed-loop systems,where sufficient conditions for the existence of controller gain matrices are formulated in terms of linear matrix inequalities(LMIs).The reachability of the proposed sliding surface can be ensured by the designed control scheme.Moreover,sufficient conditions under which the zeno behavior can be avoided is given,and the inter-execution time interval is presented.The problem of asynchronous adaptive event-triggered SMFTC for Markov jump systems with actuator failure is investigated.A hidden Markov model is introduced to describe the asynchronous phenomenon of jump modes between plants and controller.The dynamical logarithmic quantizer is applied to signal quantization both in sensorsto-controller and controller-to-actuators sides.A new event-triggered detector condition is developed based on the quantized state vector.Combined with SMC and adaptive control techniques,a new asynchronous adaptive event-triggered SMFTC law is designed,which has ability to guarantee the robust performance of the closed-loop system in the presence of the effects of actuator failures,time delay resulting by event-triggered and signal quantization error.Moreover,the developed asynchronous adaptive event-triggered SMFTC scheme can ensure that the sliding surface converges the defined small domain in finite time.The lower bound for inter-execution time is analyzed and presented,and sufficient condition under which the Zeno behaviors can be avoided is given.The problem of adaptive event-triggered SMFTC for a class of Markov jump systems with actuator failure and state-dependent attacks is investigated.By utilizing a coordinate transformation,the sliding surface is designed in the new coordinate system.The sufficient condition which can ensure the stability of sliding motion is given.To reduce the communication burden of system,a suitable event-triggered condition is developed.To ensure the stability of overall closed-loop control system and reachability of sliding surface,a new adaptive event-triggered SMFTC is proposed.Under the proposed control law,the influence of parameter uncertainties,attacks,actuator failure and state error induced by event-triggered mechanism can be completely compensated.The positive minimal inter–event time interval is estimated,which shows that the Zeno behavior is avoided.The problem of adaptive event-triggered SMFTC for Markov jump stochastic system with hybrid actuator failure and attacks is investigated.A mode-dependent sliding surface is designed,which ensure that sliding motion does not contain Brownian motion variable.Sufficient conditions for the stability of sliding motion are presented in terms of LMIs.To reduce the communication burden,a new suitable event-triggered condition is designed.A new adaptive event-triggered SMFTC strategy is proposed,which can ensure the stability of closed-loop control system and reachability of sliding surface.And under the proposed control law,the effects of external disturbance,actuator faults,hybrid attacks and state error induced by event-triggered mechanism are eliminated.The estimate of positive minimal inter–event time interval is presented,which implies that the Zeno behavior is avoided.更多还原