1. image.png

key words：

Strong influences on terminal voltage

1. Temperature
2. SOC
3. Precondition
4. Age (calendric/ cyclic)
5. HAVC Usage (Heating, ventilation, aircondition)
6. DOD, depth of charge

historical data, road profile (urban, intercity),
traffic congestion level, driver behaviour (i.e. aggressiveness),
environmental conditions (i.e. weather conditions), dynamic
vehicle parameters (i.e. SOC, SOH) and accessory loads (i.e. heating,
cooling

Many different cell
NMC\LFP\LTO

NN/ stochastic model/ Fuzzy-Logic

Pro： Little Knowledge about BMS, computational performance
Con: non-explainable, not for all batteries/ comprehensive measurement

Modeling approach

pro: computational performance, sufficient precisions, partly physical meaning
con: leaner model (U and I )

1. H.A. Yavasoglu

ML Predict Driving Range

• DT for road type
• ANN for range estimation (Use ann because the update ability)

image.png

conventional multiple linear regression method
gradient boosting decision tree algorithm

• residual usable energy (RUE)

SOC(state of charge) , SOH (state of health), temperature,
future discharge voltage
capacity values

image.png

Implementation of machine learning-based real-time range estimation method without destination knowledge for BEVs

2. Compare different algorithms Delnevo*

image.png

3 State-of-Charge Estimation Methods for Li-ion Batteries in Electric Vehicles

image.png

To estimate SOC, there are three methods

All about the SOC

image.png

image.png