Session: 08-01: Poster Session

Paper Number: 149642

149642 - Recurrence Quantification Analysis: A Quantitative Tool for Characterizing Ic Engine Combustion Cyclic Variations 

Abstract: 

Recurrence Quantification Analysis: A Quantitative Tool for Characterizing IC Engine Combustion Cyclic Variations

Muhannad Hendy, Kalyan Kumar Srinivasan, Sundar Rajan Krishnan

Department of Mechanical Engineering, The University of Alabama

 

Recurrence quantification analysis (RQA) is a powerful tool to study nonlinear system dynamics and to unveil patterns in time-series data. Previous research has showcased the use of RQA to study the behavior of dynamical systems and to characterize the onset of important system events; e.g., the seismological signatures preceding earthquakes, the identification of a misfiring engine cylinder, etc. Nonetheless, the procedure adopted for performing RQA to generate physically interpretable output has mainly been ad hoc and relies on significant experience of the analyst who employs RQA. The present work’s  objective is twofold: (1) to help demystify the main concepts of RQA and to explain the metrics in a conceptual way; (2) to propose guidelines on the determination of the parameter values used in RQA. Once the physical basis of the RQA parameter choices is understood, the interpretability of RQA results including recurrence plots (RPs) and other quantifiable metrics is enhanced. To illustrate the underlying concepts of RQA and the overall procedure used in the analysis, in-cylinder combustion pressure data obtained on a four-stroke single-cylinder research engine operating at different conditions are used. Each operating point is represented by a time-series of pressure data for 1,000 consecutive cycles with a 0.1 CAD resolution. The different operating conditions selected encompass a wide range of engine operation from extremely stable to very unstable, including both single fuel and dual fuel combustion. The coefficient of variation (COV) of the indicated mean effective pressure (IMEP) is used as a global measure of engine stability at operating conditions and the RQA results are interpreted to shed light on the system dynamics leading to the cyclic variations.  Finally, to alleviate the ad hoc nature of RQA applied to study IC engine combustion, an effective procedure is outlined for determining rational parameter choices for RQA, especially for separating the stochastic and deterministic (if present) components of the cyclic combustion variations.

Presenting Author: Muhannad Hendy The University of Alabama

Presenting Author Biography: PhD student at the department of mechanical engineering, The University of Alabama.

Authors:

Muhannad Hendy The University of Alabama
Kalyan Srinivasan The University of Alabama
Sundar Krishnan The University of Alabama