【摘要】 在现代航空发动机设计中可用于限制保护控制量的数量不断增加。为充分发挥发动机的调节能力，以变循环发动机（VCE）为被控对象，设计了基于模型预测控制（MPC）的发动机多变量限制保护器。将限制量作为状态量建立了增广的线性化模型；把限制量的极限值作为状态约束条件纳入至MPC的滚动优化过程，实现了限制保护；针对采用传统MPC方法计算量大的问题，提出了一种将限制预测模型和控制预测模型分离的设计方案，并同时引入了预测步长，有效减小了MPC的计算量。结果表明：相比于传统的基于min-max切换逻辑的限制保护控制方法，采用MPC方法的限制保护控制的性能得到了明显改善，且能完全适用于多变量系统；改进后的MPC的单位计算耗时从最长26.08 s缩短至最短0.92 s，达到传统方法的量级。更多还原

【Abstract】 In modern aeroengine design, the number of variables that can be used for limit protection continues to increase. To fully exploit the engine’s regulation capacity, a multivariable limit protection control based on model predictive control(MPC) was designed with a variable cycle engine(VCE) as the controlled object. By treating the limit values as state variables, an augmented linearized model was established, and then those limit values were incorporated into the rolling optimization process of MPC as the state constraints, achieving effective limit protections. To address the high computational load of conventional MPC methods, a design scheme was proposed that decouples the limit prediction model from the control prediction model, along with the introduction of the prediction step size, significantly reducing MPC computational demands. The results show that compared with the conventional limit protection control method based on minmax switching logic, the MPC-based method significantly improves limit protection performance and supports multivariable systems completely; the unit calculation time of the improved MPC method decreases from the highest 26.08 s to the lowest 0.92 s, reaching the level of conventional methods.更多还原