Simple Comprehensive Fuel Consumption and CO2 Emissions Model Based on Instantaneous Vehicle Power

The majority of fuel consumption models use vehicle tractive power and/or velocity as explanatory variables. These models appear to be simple and easy to implement, however they suffer from two major drawbacks, namely: (a) the use of vehicle specific power results in a bang-bang control (i.e. the optimum fuel consumption results from a full throttle acceleration) if higher orders of power are not considered in the model; and (b) the calibration of the model parameters cannot be done using publicly available data and thus requires field data collection for each vehicle. Consequently, the research presented in this paper develops two simple fuel consumption models that do not result in a bang-bang control system and that can be calibrated easily using publicly available data. Specifically, the models can be calibrated using the Environmental Protection Agency city and highway fuel economy ratings that are publicly available. The models are demonstrated to estimate vehicle fuel consumption rates consistent with in-field measurements (coefficient of determination above 0.90). Finally, a procedure for estimating CO2 emissions is developed producing emission estimates that are highly correlated with field measurements (greater than 0.98). The development of this model attempts to bridge the existing gap between traditional power-based fuel consumption models and vehicle operational control systems such as fuel-optimized cruise control systems, real-time eco-driving systems, and adaptive cruise control systems on passenger cars using road topography information.