Fuel Economy, Hypermiling, EcoModding News and Forum - EcoModder.com - Preheating saves big. Argonne tests Prius.

From here with thanks to Frank:http://www1.eere.energy.gov/vehicles...hments_rpt.pdf

Quote:

Testing Plug-in Hybrid Vehicles (PHEV) over standardized and real world drive cycles has shown relatively large fuel consumption differences between ambient cold starts and hot starts. Losses on the order of 25-40% have been observed from ambient 20oC cold starts to optimal hot temperature urban drive cycle operation (Figure 1). This is especially critical for PHEVs, when long durations between engine operation result in reduced engine temperature. Testing under this project begins to quantify these temperature effects and identifies areas for future vehicle efficiency studies and improvements to PHEVs.

Argonne researchers have applied response surface methodology techniques for PHEV experimental test data to characterize the thermal effect on efficiency [Figure 2]. Combined with a technique for predicting the engine thermal state from its initial temperature, this unique methodology accurately predicts the fuel efficiency over a drive cycle from ambient cold starts to fully operational temperature. Current work focuses on
developing a displacement-independent engine model that will allow vehicle simulations to account for engine thermal efficiency effects. Analysis from this methodology led to the following conclusions:

Engine efficiency significantly increases with increasing engine temperature.

Projected optimal engine temperature is ~25% more efficient than a 22C ambient cold start.

The initial enrichment spike encountered during a cold start accounts for a ~3% fueling increase compared to a warm engine. Greater accumulated losses (~20%) follow this cold start enrichment until the engine reaches optimum temperature.

Between the range of 25°C to 60°C, each 5°C increase in initial engine temperature decreases fuel consumption by 3.2%~1.9%, respectively. This fuel usage reduces as temperature increases, ultimately reaching an asymptote.

For the experimental PHEV, ten minute soak times may result in ~5 degree
variations in temperature (dependent upon powertrain operating temperature).

Losses associated with the electric components, rolling losses, and transaxles are minimal relative to engine and transmission thermal losses.