Stochastic model predictive control for electric vehicles thermal management

A method for controlling thermal management for electric vehicles powered in part by batteries, comprising:

• defining a cost function indicative of energy consumption by a thermal management system of a vehicle, such that the energy consumption is a function of a route taken by the vehicle, wherein the cost function accounts for multiple routes that the vehicle can take and includes slack variables and at least one input variable correlated to an input of the thermal management system;
• receiving, by a computer processor, a list of potential routes for the vehicle;
• adapting, by the computer processor, penalty weights for slack variables in the cost function based on predicted constraint violations;
• calculating, by the computer processor, a thermal management cost by minimizing the cost function using dynamic programming, thereby determining a value for the at least one input variable;
• and
• controlling the thermal management system in accordance with the value for the at least one variable.

A thermal management system for an electric vehicle powered in part by batteries, comprising:

• a cooling subsystem including a battery cooling loop and a cabin cooling loop;
• a pump configured to pump refrigerant through the cooling subsystem;
• a value fluidly coupled between an outlet of the pump and the battery cooling loop and the cabin cooling loop, such that the valve controls amount of refrigerant flowing into the battery cooling loop and the cabin cooling loop;
• a data store storing a cost function indicative of energy consumption by the thermal management system, such that the energy consumption is a function of a route taken by the vehicle, wherein the cost function accounts for multiple routes that the vehicle can take and includes at least one input variable correlated to an input of the thermal management system;
• and
• a controller interfaced with at least one of the pump and the valve, the controller configured to receive a list of potential routes for the vehicle and calculate a thermal management cost by minimizing the cost function, thereby determining a value for the at least one input variable.