Machine learning based EV vacuum pump on/off strategy optimization tool

Vacuum pump in a passenger car is used to increase the comfort and safety of the driver by providing assistance to braking. In a convectional vehicle, vacuum pump is coupled to alternator or to engine, as a result of this vacuum pump runs at speed of the engine and supplies continuous vacuum to the brake booster irrespective of the vacuum requirement. In Electric vehicles, due to absence of Engine, use of motor driven vacuum pump is must to supply vacuum. With suitable electronic control, the motor could be turned on only when required thus saving a significant energy. But the only drawback of electric vacuum pump is, it is not having provision for cooling like in conventional vacuum pump which is oil cooled. Vehicle manufactures optimize Vacuum pump ON/OFF duty cycle. In this work, a model is developed using Machine Learning to optimize the electric vacuum pump ON/OFF strategy with respect to customer comfort, safety and derive the pump life by avoiding vehicle level tuning and the developed model is validated by correlating the obtained results with field test results.