GOALI: HCCI Engine Control and Optimization Using Extremum Seeking

随着减少排放的压力越来越大，对新的低污染发动机技术的需求从未如此之大。因此，具有类似柴油机的效率、低微粒和低NOx排放的HCCI发动机是一个主要的候选者。然而，由于HCCI发动机没有直接的梳状燃烧触发器，因此控制梳状燃烧的开始是一项复杂的任务。为了使这项技术成为现实，我们已经组建了一个具有HCCI发动机基础知识背景的团队（La Lawrence Wrence Li Livermore National Labs LLNL）、最先进的汽车控制和发动机映射（F福特，ord）以及先进的非线性控制方法（UCSD），我们计划解决HCCI发动机映射问题以及依赖于发动机映射的发动机控制。由于需要调整的发动机参数很多，因此HCCI发动机的映射是一个复杂且耗时的过程。我们计划在这个问题上采用的寻优方法（ES）已经在福特奥德汽车公司得到了验证，它是一种有效的工具，相对于传统的技术，它可以将火花点火发动机的映射时间缩短三到五倍，在映射过程中可以节省几个月的时间。同样的好处可以预期与HCCI发动机。本文将对福特的算法进行改进，并在常规发动机和实验性HCCI发动机上测试这种新方法，由于HCCI发动机对温度的敏感性，除了映射和优化之外，HCCI发动机在运行过程中还需要反馈控制。我们将构造非线性控制算法来控制梳状燃烧使用过程，建立在我们以前的控制气体作为涡轮机燃烧室的结果。我们的算法将被应用到具有详细化学动力学的HCCI发动机模型以及LLNL和福特的实验性发动机上。这些将是使HCCI发动机技术达到一个较低水平的有益步骤，在这个水平上，我们可以受益于其优于传统内燃机的优点。这项研究将有助于使HCCI发动机技术成为现实，这是非常重要的，因为他们的承诺，绿色运行和日益严格的排放标准。该程序将使专家系统算法在多变量、变量、约束和快速在线优化方面得到进一步的发展。这些进步将应用于许多其它技术。

With ith increasing pressure to reduce emissions, the need for ne new lo low polluting engine technologyhas ne never er been greater greater. Consequently Consequently, the HCCI engine, with diesel lik like ef efficienc ficiency, yet lo low par par-iculate and NOx emissions, is a prime candidate. Ho However er, because HCCI en engines gines do not ha have adirect comb combustion ustion trigger trigger, controlling the start of comb combustion ustion is a complicated task. To help mak makethis technology a reality we ha have assembled a team with a background in HCCI engine fundamen- fundamentalstals (La Lawrence wrence Li Livermore ermore National Labs LLNL), state of the art automoti automotive control and enginemapping (F Ford), ord), and adv advanced anced nonlinear control methodologies (UCSD).We plan on addressing HCCI engine mapping issues, as well as engine control, which relieson engine maps. Mapping HCCI engines is a complicated and time consuming process due to theman many engine parameters which must be tuned. Extremum seeking (ES), which we plan to emplo employfor this problem, has been demonstrated at Ford ord Motor Compan Company to be an ef effecti fective tool to reducethe time to map spark ignition engines by a factor actor of three to fi five relati relative to traditional techniques,which can sa save months during the mapping process. The same benefits can be expected xpected with HCCIengines. We wil will impro improve upon Ford' ord's already successful algorithm and test this ne new method oncon conventional entional engines and an experimental xperimental HCCI engine.In addition to mapping and optimization, HCCI engines require feedback control during op- operationeration due to their sensiti sensitivity vity to temperature. We will construct nonlinear control algorithms tocontrol the comb combustion ustion process, building uilding upon our prior results on control of gas as turbine enginecomb combustors. ustors.Our algorithms will be applied to models of the HCCI engine with detailed chemical kineticsand on experimental xperimental engines at LLNL and Ford. ord. These will be helpful steps to bring the HCCIengine technology to a le level el in which we can benefit from its adv advantages antages over er con conventional entional internalcomb combustion ustion engines.Impact Statement. This research will help mak make HCCI engine technology a reality reality, which isimportant due to their promise of green operation and increasingly stringent emissions standards.Intellectual Merit. The program will mak make further adv advancements ancements to ES algorithms, especially inmulti multivariable, ariable, constrained, and fast ast on-line optimization. Such adv advances ances will ha have application toman many other technologies.