针对160MW锅炉-汽轮机系统的多目标控制问题和计算负担，提出一种事件触发多目标经济模型预测控制(multi-objective economic model predictive control，MO-EMPC)策略。结合字典序方法和生产过程的需求，将锅炉-汽轮机补给燃料的调节控制作为最高优先级控制目标，主蒸汽调控作为第2层优先级控制目标，给水阀调控作为第3层优先级控制目标，构建分层滚动时域优化控制问题。引入锅炉-汽轮机经济最优平衡点定义辅助正定函数的最优值函数，并利用该最优值函数设计收缩约束条件，通过松弛收缩约束条件设计触发机制，建立具有稳定性和经济性能最优的轻量化多目标预测控制方案。对比时间触发下的字典序MO-EMPC算法，所提方法不仅通过降低求解频次减少了计算量，同时保证了锅炉-汽轮机系统的经济性能。
Event-triggered multi-objective economic predictive control of boiler-turbine systems
For the multi-objective control problem and computational burden of 160 MW boiler-turbine systems,this paper proposes an event-triggered multi-objective economic model predictive control(MO-EMPC)strategy.Combining with the lexicographic method and the production process requirements,boiler refueling is taken as the highest priority control objective,the main steam control is taken as the second priority control objective,and feed valve control is taken as the third priority control objective.A hierarchical receding horizon optimal control problem is formulated accordingly.The auxiliary positive-definite functions at boiler-turbine systems economic optimal equilibrium points are defined,which are used to design the contraction constraint condition.The triggered mechanism is then designed by relaxing contraction constraint.A lightweight multi-objective predictive control scheme with guaranteed stability and economic performance is established.Compared with the time-triggered lexicographic MO-EMPC,the proposed method decreases the calculation amount by reducing the solution frequency,and realizes the economic performance of the boiler-turbine system.
boiler-turbine systemevent-triggered controleconomic model predictive controlmulti-objective controlstability
万方数据
维普
