CyberGIS for Geospatial Discovery and Innovation by Shaowen Wang & Michael F. Goodchild

Author:Shaowen Wang & Michael F. Goodchild
Language: eng
Format: epub
ISBN: 9789402415315
Publisher: Springer Netherlands

4.2.4 Improving Physics-Based Models via Novel SBD Patterns

Every engine or vehicle phenomenon (e.g., engine combustion) is explainable by physical laws; however, existing lean predictive models minimize computation by utilizing only a few measurement parameters, thereby limiting the possibility to accurately predict all possible phenomena (e.g., reduced accuracy when predicting vehicle emissions). SBD patterns, from engine measurements data, combined with physics-based phenomenological models, can be constructed to accurately predict NO and CO emissions. Such models are useful for assisting engine control and after-treatment systems for the engine emissions.

Opportunity 4: Improving physics-based emission models via novel SBD patterns. SBD of engine measurements creates an opportunity for studying combustion and engines in real-world scenarios under the effect of traffic, road conditions, weather and driver behavior which can lead to enhancing the physics-based emission models and thus lead to better vehicle design with improved fuel efficiency and emissions. Figure 9 shows a framework that illustrates how spatio-temporal engine measurement data can be used to better understand and improve engine and combustion science. First, in-coming data collected from vehicles in the real-world is tested for divergence in emissions or fuel economy based on regulatory emission standards, time and geographic location, etc. If a vehicle or fleet of vehicles exhibit divergence, there exist two potential pathways for understanding the problem: structured patterns or unstructured patterns. Structured or internal patterns are those that arise from known physics-based models or rules by using the available recorded vehicle data. Unstructured or external patterns must be studied further through mining engine measurement data and identifying novel engine patterns (e.g., linear emission hotspots along the road network and engine signatures that co-occur within these hotspots) to find correlations that are not immediately apparent from the physics-based model using recorded ambient data external to the physics-approach. Should statistically significant patterns emerge and survive evaluation in engine laboratory, model improvements can then be recommended to develop an understanding of vehicle operation. Once physical understanding of a divergence event is determined, the information can either be used in the engineering of future vehicles or to enable cloud connected vehicle adaptation in real-time.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.