Online optimization of engine control unit to satisfy performance and drivability metrics

Abstract

Purpose: The purpose of this study is to develop combined objective evaluation metrics for drivability and performance attributes and a self -tuning algorithm that yields optimized performance and drivability based on developed metrics throughout the life span of the vehicle over against changes due to wear and aging. Theory and Methods: Firstly, to demonstrate method performance, tip -in during deceleration maneuver is selected, and event detection conditions are defined. For assessments, drivability and performance criteria are determined. Design of experiments is prepared and performed with small passenger and light commercial vehicles. For subjective assessments, each criterion is rated separately and is done by means of drive rating from 1 to 10 based on developed rating scale. Response surface methodology is adopted to correlate objective data and subjective assessments, and rating functions for each criterion are formulated. For combined objective evaluation, overall tip -in rating is also formulated from criteria ratings using correction factors for each rating and weight factors of each criterion. To demonstrate effectiveness of the proposed method optimization is studied online on a test vehicle. Golden section algorithm, and a multiplier value that updates the look -up table which defines torque rate based on instantaneous torque is used for optimization. The proposed methodology is applied and validated in the vehicle. Results: For models developed for criteria functions, root mean square error of residuals are all below 0.2531 (between 0.1139 and 0.2531) and coefficient of determination values are all above 0.9252 (between 0.9901 and 0.925) which shows strong relationship between objective measurements and subjective assessments. Furthermore, ratings from developed tip -in rating function and experts' subjective results are within +/-0.1 rating range which represents the validity of the optimization process. Conclusion: The methodology has been proven to be successful both in vehicle evaluations and in providing online performance and drivability optimization. The recommended software structure is suitable for diagnostics purposes to track changes in system operation and avoid system failures. In addition, the proposed methodology can reduce vehicle testing time and development costs, and it is easy to apply without any expertise. Moreover, it has high potential to be used in an engine control unit due to its simplicity.