Control of quasi-homogeneous combustion in diesel engines using fully-variable injection

这个子项目的主要目标是使用基于优化的控制来进一步发展预混压燃(PCCI)柴油燃烧过程。为此，整个喷射过程和整个燃烧函数分别用作连续驱动变量和控制变量。为了能够建立周期间控制，理想情况下也是周期内控制，这些变量必须用基于物理的低维参数化来描述。在第一个资金期，开发了基于无监督和有监督机器学习相结合的定量模型，并根据各种多次喷射模式的实验数据进行了验证，其中包括所谓的具有极限气缸压力升高率的阿尔法过程。利用TP6的数据，建立了反应动力学模型，作为三维CFD模拟和模型简化的基础。为了在较宽的负荷范围内甚至在过渡工况下实现对各种性能参数的控制，将采用低温燃烧(LTC)和阿尔法过程相结合的方法，并将建立循环控制方法。这一点的基础是建立一个快速而精确的模型来描述早期放热对性能参数的影响。首先利用实验获得的数据空间分析LTC与极限气缸压力升高率之间的相互作用及其对性能参数的影响。这形成了通过结合基于反应动力学的半物理建模方法来描述喷射/燃烧功能和燃烧功能/排放关系的基础。为了在实时燃烧控制中的应用，这些都将转化为一个低维的参数化。在这里，压力曲线分析用于描述LTC的唯一性将被严格地检验，并探索离子电流测量的电势作为补充参数，甚至替代压力曲线分析。基于反应动力学的敏感度的非常早期点火阶段，这是非常重要的低温燃烧，因为不断增加的时间比例的化学点火延迟。将为该项目和其他子项目开发动力学模型，其中还包括离子流和污染物形成机制，并例如在降阶模型中使用。CFD模拟将被用来识别和解释相关性，并将推导出相应的物理关系来支持简化模型。将与TP6合作开展LTC的详细动力学、污染物排放和离子电流表征方面的工作。简化的模型将与TP7和单缸研究发动机合作进行验证。

The main goal of this subproject is the further development of the Premixed Charge Compression Ignition (PCCI) diesel combustion process using optimization-based control. For this, the entire injection process and the entire burn function are used as continuous actuation and control variables, respectively. In order to be able to establish a cycle-to-cycle, ideally also an in-cycle control, these variables must be described by physically-based, low-dimensional parameterizations. In the first funding period, quantitative models based on a combination of unsupervised and supervised machine learning were developed and validated from experimental data for various multiple-injection patterns including the so-called alpha-process with limiting cylinder pressure rise rate. With data from TP6, models for reaction kinetics were developed that serve as the basis for 3D CFD simulations and model reduction. In order to apply the control of various performance parameters over a wide load range and even for transient operation conditions, a combination of low-temperature combustion (LTC) and alpha-process will be implemented, and an in-cycle control approach will be established. The basis for this is a fast and precise model to describe the influence of the very early heat release on the performance parameters. The analysis of the interaction between LTC and limited cylinder pressure rise rates as well as its effect on the performance parameters will first be carried out using experimentally obtained data spaces. This forms the basis for the description of the relationships injection/burn function and burn function/emissions by incorporation of semi-physical, reaction-kinetics based modeling approaches. For the application in real-time combustion control, these will be transformed into a low-dimensional parameterization. Here, the uniqueness of the pressure curve analysis for the description of the LTC will be critically examined and the potential of ion current measurements explored as a complementary parameter or even to substitute the pressure curve analysis. The reaction-kinetics based sensitivities of the very early ignition phase will be examined, which are of great importance in low-temperature combustion due to the increasing temporal proportion of the chemical ignition delay. Kinetic models, which also include the ion current and pollutant formation mechanisms, will be developed for this and other sub-projects, and used, for instance, in the reduced order models. CFD simulations will be used to identify and explain correlations and corresponding physical relationships will be derived to support the simplified model. Work on the detailed kinetics of LTC, pollutant emissions, and ion current characterization will be carried out in collaboration with TP6. The simplified models will be validated in cooperation with TP7 and with the single-cylinder research engine.