US 11,704,590 B2
Methods and systems for predicting failure of a power control unit of a vehicle
Shailesh Joshi, Ann Arbor, MI (US); Hiroshi Ukegawa, Northville, MI (US); Ercan M. Dede, Ann Arbor, MI (US); and Kyosuke N. Miyagi, Ann Arbor, MI (US)
Assigned to Toyota Motor Engineering & Manufacturing North America, Inc., Plano, TX (US)
Filed by Toyota Motor Engineering & Manufacturing North America, Inc., Erlanger, KY (US)
Filed on Mar. 24, 2017, as Appl. No. 15/468,618.
Prior Publication US 2018/0276546 A1, Sep. 27, 2018
Int. Cl. G05B 13/04 (2006.01); G06N 20/00 (2019.01); G05B 13/02 (2006.01)
CPC G06N 20/00 (2019.01) [G05B 13/026 (2013.01); G05B 13/0265 (2013.01); B60Y 2200/91 (2013.01); B60Y 2200/92 (2013.01); B60Y 2306/13 (2013.01); Y10S 903/903 (2013.01)] 19 Claims
OG exemplary drawing
 
1. A method for predicting a failure of a power control unit of a vehicle, the method comprising:
obtaining data including a simulated voltage, a simulated current, and a simulated temperature of the power control unit from a plurality of sensors configured to measure the simulated voltage, the simulated current, and the simulated temperature of the power control unit of the vehicle being subject to simulated multi-load conditions including a simulated thermal cycling, a simulated power cycling, and a simulated vibration cycling simultaneously, the power control unit being configured to control power from a battery of the vehicle, the power control unit being a separate element from the battery;
implementing a machine learning algorithm on the data including a temperature of the power control unit to obtain machine learning data;
obtaining new data including a real voltage, a real current, and a real temperature of the power control unit from the plurality of sensors for measuring the power control unit of the vehicle being subject to real multi-load conditions including a thermal cycle, a power cycle, and a vibration;
implementing the machine learning algorithm on the new data to obtain test data;
predicting that the power control unit is going to fail based on a comparison between the test data and the machine learning data, and
sending, to a brake electronic control unit, a command for adjusting a regeneration required torque in advance of a failure of the power control unit in response to predicting that the power control unit is going to fail.