Characterization of knock intensity distributions: Part 1: statistical independence and scalar measures

Spelina, Jill M.; Jones, James C. Peyton; Frey, Jesse

doi:10.1177/0954407013496233

Cited by 57 publications

(48 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While it is possible that degree of cyclic or cylinder-to-cylinder correlation may increase at other conditions due to physical feedback mechanisms such as latent heat from extreme knock, high EGR, or valve timings that give high levels of residual gas, these results suggest that to a reasonable first approximation, the data behaves as a cyclically independent random process on both engine platforms considered. Such a conclusion is also consistent with a more extensive study of cyclic independence on a different engine across a large database of test results, [28].This is important because it means that the knock intensity of the forthcoming cycle is the outcome of an independent random event which by definition cannot be controlled directly. Standard control theory and methods therefore cannot be applied, and some form of stochastic control is necessary in order to control some statistical measure of the distribution from which knock intensities are drawn.…”

Section: Methodssupporting

confidence: 63%

See 1 more Smart Citation

Experimental Validation of a Likelihood-Based Stochastic Knock Controller

Thomasson

Shi

Lindell

et al. 2016

IEEE Trans. Contr. Syst. Technol.

View full text Add to dashboard Cite

Abstract-New likelihood-based stochastic knock controllers have the potential to deliver significantly improved regulatory response relative to conventional strategies, while also maintaining a rapid transient response, but evaluation studies to date have only been performed in simulation. In this paper an experimental validation of the new strategy is presented. To demonstrate the robustness of the method, the algorithm is implemented on two different engine platforms, using two different knock intensity metrics, and evaluated under different operating conditions. One of these platforms is a five cylinder variable compression ratio engine, enabling the controller to be tested under different compression ratios, as well as different speed and load conditions. The regulatory and transient performance of the likelihood-based controller is assessed in a back-to-back comparison with a conventional knock controller and it is shown that the new controller is able to operate closer to the knock limit with less variation in control action without increasing the risk of engine damage.

show abstract

Section: Methodssupporting

confidence: 63%

“…3. Although it is widely acknowledged that the process is random, it is also important to check whether the data is cyclically uncorrelated, [28]. A correlation analysis was therefore performed on the data from both engines based on 200 cycles of data.…”

Section: Methodsmentioning

confidence: 99%

Experimental Validation of a Likelihood-Based Stochastic Knock Controller

Thomasson

Shi

Lindell

et al. 2016

IEEE Trans. Contr. Syst. Technol.

View full text Add to dashboard Cite

show abstract

“…To a reasonable approximation, therefore, the process may be regarded as a cyclically independent process, at least on this engine and within the BL 2 4°... BL+3°range considered. Further details and analysis may be found in Spelina et al 25 The assumption of cyclic independence has a profound impact on knock control systems since it means that the apparent control variable, knock intensity, is a random variable that cannot be predicted or controlled directly. Instead, it is only possible to control some statistical aspect of the knock process or distribution from which these intensities are drawn.…”

Section: Properties Of the Knock Processmentioning

confidence: 99%

Stochastic simulation and analysis of a classical knock controller

Spelina

Jones

Frey

2014

International Journal of Engine Research

Self Cite

View full text Add to dashboard Cite

Knock control remains a critical issue in modern engine powertrains, and a renewed emphasis on knock as a stochastic process has proved beneficial in the development of new controller designs. However, the random nature of knock also makes it hard to evaluate the closed-loop performance of a knock controller in a rigorous, repeatable way. This work therefore focuses particularly on the statistical properties of knock intensities and knock events, and a new Markovbased analysis is used to compute the corresponding statistical properties and distribution of the closed-loop response. The method is applied to a conventional knock controller, revealing new aspects of its behavior. In particular, the closedloop spark advance distribution is found to be periodic initially, only collapsing to an invariant steady-state distribution as a result of limits applied to the spark advance actuation. The stochastic response of the controller to different initial conditions is also investigated, providing a more rigorous insight into its performance. The results of the Markov-based analysis are confirmed using Monte Carlo simulations.

show abstract

“…Prior work has shown that knock intensity and knock events have characteristic statistical properties that can be leveraged for control purposes [9][10][11][12]. Spelina et al [10,11] explored and quantified the statistical characteristics of the knock process across a broad range of engine operating conditions. Their studies showed that, for the operating points they studied, knock behaved to a good first approximation as a cyclically independent random process.…”

Section: Introductionmentioning

confidence: 99%

Knock Intensity Distribution and a Stochastic Control Framework for Knock Control

Kassa

Hall

Pamminger

et al. 2019

Journal of Dynamic Systems, Measurement, and Control

View full text Add to dashboard Cite

One of the main factors limiting the efficiency of spark-ignited engines is the occurrence of engine knock. In high temperature and high pressure in-cylinder conditions, the fuelair mixture auto-ignites creating pressure shock waves in the cylinder. Knock can significantly damage the engine and hinder its performance; as such, conservative knock control strategies are generally implemented that avoid such operating conditions at the cost of lower thermal efficiencies. Significant improvements in the performance of conventional knock controllers are possible if the properties of the knock process are better characterized and exploited in knock controller designs. One of the methods undertaken to better characterize knocking instances is to employ a probabilistic approach, in which the likelihood of knock is derived from the statistical distribution of knock intensity.

show abstract

Characterization of knock intensity distributions: Part 1: statistical independence and scalar measures

Cited by 57 publications

References 22 publications

Experimental Validation of a Likelihood-Based Stochastic Knock Controller

Experimental Validation of a Likelihood-Based Stochastic Knock Controller

Stochastic simulation and analysis of a classical knock controller

Knock Intensity Distribution and a Stochastic Control Framework for Knock Control

Contact Info

Product

Resources

About