基于自适应延时PI控制器的汽油机λ双环控制策略的研究

Study on λ Dual-Loop Control Strategy for Gasoline Engines Based on PI Controller with Adaptive Time Delay

针对三元催化器前后都配置了开关型氧传感器的汽油发动机,提出了一种新型的双闭环λ自适应控制策略。前氧传感器用于对闭环控制因子进行PI反馈控制,后氧传感器用于实时监测并通过冻结PI控制修正λ偏差。如果前氧传感器老化或中毒,电压输出特性会发生明显漂移,传统PI闭环控制无法准确控制λ为1左右,导致三元催化器转换效率下降。根据后氧传感器电压偏差计算冻结模式和冻结时间,在前氧传感器信号跳变时刻冻结PI控制,使闭环控制因子延时反跳。在冻结时间内闭环控制因子维持在极大/极小值状态,利用单边时间延迟效应构成不对称PI控制模式精确补偿λ漂移,确保λ被控制在理想运行范围,保证最优的三元催化器转换效率。利用先进的ASCET开发环境进行图形化建模并自动生成C代码,与底层驱动代码集成后,下载到32位MPC5634硬件平台安装到发动机台架上进行试验。研究结果表明:当前氧传感器失常时,λ仍然可以被精确控制在0.99~1.01的理想范围内。

For an engine equipped with dual switch-type oxygen sensors upstream and downstream of the three-way catalyst（TWC） , an innovative adaptive dual-loop λ control strategy was proposed. The front oxygen sensor upstream of the TWC was used to perform PI feedback control of the closed-loop control factor and the rear sensor downstream of the TWC was used to monitor and correct the A in real-time. If the front oxygen sensor voltage output characteristic shifts remarkably due to aging or oil-poisoned？ the traditional closed-loop PI method can not prevent the A from deviating from 1 , resulting in a deteriorated efficiency of the TWC. Therefore, in the proposed control strategy, every time the front oxygen sensor voltage steps, a freezing mode and a freezing period are calculated depending on the offset of the rear oxygen sensor voltage from its reference value and applied to freeze PI adjustment and delay the closed-loop control factor to invert jump for a specific period. The closed-loop control factor is maintained at the maximum/minimum state during the freezing period. In this way, an asymmetric controller with unilateral time delay is formed to compensate the A offset so that the A can be controlled precisely in the ideal operation range to achieve an optimum conversion efficiency of the TWC. The advanced software ASCET is used to graphically model the controller and automatically generate C codes. After being integrated with the baseline driver code, C codes are downloaded to MPC5634 hardware platform and installed on an engine bench to experimentally demonstrate the effectiveness of the proposed control strategy. The results show that the A can still be precisely controlled within the ideal range of 0. 99 to 1. 01 even if the front oxygen sensor malfunctions.