Tesla Battery Health: Surprising Impact of Model Year

Average predicted State of Health (SOH) for Tesla vehicles by model year, based on standardized odometer readings (10,000, 50,000, 100,000, and 200,000 miles). Note the notably higher SOH from 2021 onward, potentially influenced by battery replacements in the dataset, and the linear trend indicating consistent battery improvements across newer models.

During the recent development of a Deep Neural Network (DNN) for predicting State of Health (SOH) and detecting abnormal battery conditions using various variables, we became curious about how battery degradation in Tesla vehicles is influenced by their production year. To explore this, we conducted a simplified additional analysis by building a basic DNN model using only the vehicle’s model year and odometer reading as inputs to predict SOH.

To isolate the influence of model year from the effect of mileage, we predicted SOH at standardized odometer readings of 10,000, 50,000, 100,000, and 200,000 miles.

The resulting graph clearly illustrates the average predicted SOH according to the model year. Interestingly, Tesla vehicles from 2021 exhibit noticeably higher SOH compared to older models, likely due to the inclusion of vehicles with replaced batteries in our training dataset.

Surprisingly, contrary to our initial expectations, the predicted SOH shows a nearly linear increase with newer model years. This finding suggests that, in addition to mileage, the production year of the vehicle has a significant impact on battery health. It also highlights the importance of proper battery management, even during periods when the vehicle is not in use.

Additionally, going forward, Dr.EV will incorporate both DNN-predicted SOH and AI-based anomaly detection.