Systems and methods include a method for optimizing real-time energy use, including balancing steam and condensate. Real-time equipment readings are received from plural pieces of equipment at a facility. A steam and condensate balancing model is executed using the real-time equipment readings. The steam and condensate balancing model uses specialized optimizer engine code to balance steam output and condensate output. A set-point optimization model is executed with selection optimization turned off to identify optimum values for boilers, STGs, letdowns, and deaerators. The set-point optimization model is executed with selection optimization turned on to identify optimized settings for drivers of turbines and motors. Setting updates are provided to the plural pieces of equipment based on the executing.