Livelocks, like deadlocks, can result in serious results in running process of flexible manufacturing systems(FMSs). Current deadlock control policies(DCPs) based on mixed integer programming(MIP) cannot detect siphons that cause and cope with livelocks in Petri nets. This study proposes a revised mixed integer programming(RMIP) method to directly solve the new smart siphons(NSSs) associated with livelocks in a system of sequential systems with shared resources(S^4 R), a typical subclass of generalized Petri net models. Accordingly,the solved NSSs are max'-controlled by adding the corresponding control places(CPs). As a result, an original S^4 R system with livelocks can be converted into the live controlled Petri net system. The related theoretical analysis and an example are given to demonstrate the proposed RMIP and the corresponding control algorithm(CA).
LI ShaoyongWU ZongliLI HailinWEI XianhongCAI YingHOU Caiqin
汽—水型热交换器作为热力站当中的重要设备之一,具有非线性、参数时变和时滞等动态特性。首先,综合分析集中供热工艺和自动控制的相关要求,对汽—水型热交换器的各个组成单元分别进行建模,确定主控制器为供水温度PIλDμ控制器(Water Supply Temperature Fractional Order Proportional Integral Derivative Controller, WST-FOPIDC)和副控制器为蒸汽流量PIλ控制器(Steam Flow Fractional Order Proportional Integral Controller, SF-FOPIC)的串级控制策略。其次,应用修正的粒子群优化算法(Modified Particle Swarm Optimization Algorithm,MPSOA),分别整定出WST-FOPIDC和SF-FOPIC的8个控制器参数最佳值。最后,基于MATLAB工具,对该串级系统进行组态和数值仿真相应地控制效果。结果表明,该串级系统在理论上是可行的,可以获得更为满意的控制效果,且供水温度的控制效果明显优于基于Ziegler-Nichols(Z-N)整定法的供水温度PID-蒸汽流量PI串级调节系统。
鉴于蒸汽压缩式制冷机组蒸发温度T_(e)与过热度D_(sh)的控制回路之间存在强耦合及大惯性、非线性和时延等特性。提出了一种蒸发温度与过热度的前馈解耦PID控制策略,且设计出改进多目标人工鱼群算法(Modified Multi-objective Artificial Fish Swarm Algorithm,MMOAFSA)对相应的PID控制器参数进行整定,以提升T_(e)与D_(sh)的调节质量。首先,对两个控制环路:电子膨胀阀开度O_(EEV)—蒸发温度T_(e)和压缩机驱动电机的供电频率f—过热度D_(sh),通过前馈补偿解耦方式来消除这两个控制回路之间的耦合效应。其次,对基本型单目标人工鱼群算法的视野V和步长S进行指数递减变化,构建改进单目标人工鱼群算法(Modified Single Objective Artificial Fish Swarm Algorithm,MSOAFSA)。再将多目标优化的混沌局部搜索策略引入MSOAFSA,设计了MMOAFSA。考虑绝对积分时间误差(Integrated Time Absolute Error,ITAE)、调节时间tc和稳态误差绝对值Ess,将min(ITAE,tc,Ess)作为MMOAFSA的多目标适应度函数,并应用该MMOAFSA对两个控制器的6个参数(KP_(1),KI_(1),KD_(1),KP_(2),KI_(2),KD_(2))进行多目标寻优,获取了相应的Pareto最优解。最后,借助MATLAB工具,对VCRU双参数前馈解耦PID控制系统(Two-Parameter Feedforward Decoupling PID Control System for VCRU,VCRU-TPFDPIDCS)组态与数值模拟。结果表明:该控制策略能够消除控制回路之间的耦合效应,同时MMOAFSA对两个控制器6个参数的自适应整定是可行的,且对T_(e)与D_(sh)的调节质量也明显优于传统的PID调节方式。
