A Bayesian Knock Event Controller

2023

Self Cite

Classical knock control strategies are still widely used in production vehicles but exist in several variant forms whose performance with respect to one another, and to engine efficiency, ease of calibration and analytical tractability has not been rigorously evaluated. This paper examines the absolute and relative performance of these strategies subject to both transient and systematic disturbances using both time history simulation methods and analytical methods. The former illustrates the response to specific instances of the knock process, while the latter enables the statistical properties of the closed loop response to be computed and compared.

Section: Retard-only Strategymentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

A comparative study of fixed-gain classical knock control strategies

2023

Self Cite

“…Despite a growing number of advanced event-based [1][2][3][4] or intensity-based [5][6][7][8][9] knock control system designs, the majority of production vehicles still employ some variant of a classical knock control strategy where the spark is retarded sharply whenever knock events occur, or advanced gradually after non-knocking cycles. These strategies are simple and robust, but the relative merits and demerits of the different variants have only recently begun to be investigated.…”

Section: Introductionmentioning

confidence: 99%

“…Indeed, a fully Bayesian knock controller has been proposed based on this principle. 4 Severity-based classical knock control therefore has some potential merit, but the benefit (if any) relative to fixed gain classical controllers has not been critically evaluated. Such evaluations are complicated because the closed loop response of the system depends on the specific instance of random knock events that are observed.…”

Section: Introductionmentioning

confidence: 99%

A critical analysis of classical severity-based knock control strategies

2023

Self Cite

A common variant of classical knock control strategies replaces fixed magnitude control moves with control actions that depend on the knock severity of the prior cycle. Though appealing to notions of proportional control, it is unclear whether such strategies offer benefit given the random nature of knock events. In this paper a new generalized event-based framework is developed and used to extend both time domain simulation and stochastic analysis methods for analyzing the closed loop behavior of such systems. These tools are then used to perform a rigorous analysis and comparison between standard advance-retard and severity-based control strategies. The analysis is also repeated using a lower knock threshold definition. The results show that the performance of severity-based knock control differs minimally from a similarly tuned standard advance-retard controller, and there is little benefit to recompense the additional complexity of severity-based control.

“…1,2 Although this is a well-researched field, recent work has shown that adopting a stochastic framework gives rise to new analytical and predictive tools, as well as advanced knock controller designs. 3–7 However, the performance of any knock control system depends fundamentally on the knock “detection” method used to reduce a knock signal time history to a scalar knock intensity metric for each cycle. Despite a very extensive literature on this topic, 8–12 relatively little attention has been given to the stochastic properties of the raw knock signals from which the scalar knock intensities are derived.…”

Section: Introductionmentioning

confidence: 99%

The statistical characteristics of knock signal waveforms

2021

Self Cite

Individual instances of the knock resonant response are easy to acquire but these are subject to noise and vary considerably from cycle to cycle due to random variations in the knock process. This work provides a new way to quantify and model the stochastic properties of knock signals, capturing both the time domain resonant characteristics within a cycle as well as the random variations from cycle to cycle. A new phase alignment method enables the ensemble mean knock waveform to be identified from the data which also removes noise components without the need for narrowband filtering. This ensemble waveform shows the empirical characteristics of knock onset, decay, and frequency slurring within the cycle as the gas expands and cools. The phase-aligned cyclic variations of the knock waveform are also shown to approximate a (time-varying) dual-Gaussian distribution, and fitting such a model to the data enables the statistical properties of the dataset as a whole to be decomposed into separate knocking and non-knocking populations providing further insight into the knock process. The technique is applied both to filtered cylinder pressure signals and to accelerometer-based knock signals, and the results are compared and contrasted.