Demand charge minimization in behind-the-meter energy management systems

1. A method for controlling behind-the meter energy storage/management systems (EMSs) for battery-optimized demand charge minimized operations, comprising:

determining an optimal monthly demand charge threshold based on at least one of a received customer load profile and a customer load profile and savings, the determining further comprising:

iteratively performing daily optimizations to determine a daily optimal demand threshold for each day of a month;

selecting a monthly demand threshold by clustering the daily optimal demand thresholds for each day of the month into one of a plurality of groups, and

determining a dominant group representative of a load pattern for a next month; and

determining a mean demand threshold for the dominant group, and

selecting the mean demand threshold as the monthly demand threshold; and

providing continuous battery-optimized demand charge minimized EMS operations based on the monthly demand threshold using a real-time controller configured for overriding the optimal charging/discharging profiles when a monthly demand threshold violation is detected.

2. The method as recited in claim 1 , wherein the clustering is executed using k-means unsupervised learning.

3. The method as recited in claim 1 , further comprising performing monthly demand threshold optimization based on monthly demand charge, average battery state-of-charge (SoC), and accumulated battery depth-of-discharge (DoD).

4. The method as recited in claim 1 , further comprising modeling and controlling daily battery utilization, charging actions, and discharging actions as defined by the monthly demand threshold (DT), power demand by a consumer, and battery limits.

5. The method as recited in claim 1 , wherein the plurality of groups from the clustering include distinct groups for weekends, weekdays, and events.

6. The method as recited in claim 1 , wherein data utilized for the daily optimizations includes one or more of maximum power demand, average power demand, daily demand threshold, demand charge saving, battery discharged energy, and average battery state-of-charge (SoC).

7. The method as recited in claim 1 , wherein the optimal monthly demand charge threshold selection is based only on a received customer load profile, the load profile including features of maximum power demand and average power demand.

8. The method as recited in claim 1 , wherein the optimal monthly demand charge threshold selection is based on a received customer load profile and savings, the load profile and savings including features of maximum power demand, average power demand, demand threshold, demand charge saving, average battery state-of-charge (SoC), and accumulated discharged energy.

9. A system for controlling behind-the meter energy storage/management systems (EMSs) for battery-optimized demand charge minimized operations, comprising:

a processor device operatively coupled to a non-transitory computer-readable storage medium, the processor being configured for:

determining, using a monthly layer constructor, an optimal monthly demand charge threshold based on at least one of a received customer load profile and a customer load profile and savings, the determining further comprising:

iteratively performing, using a demand threshold optimizer, daily optimizations to determine a daily optimal demand threshold for each day of a month;

selecting a monthly demand threshold by clustering the daily optimal demand thresholds for each day of the month into one of a plurality of groups, and determining a dominant group representative of a load pattern for a next month; and

determining a mean demand threshold for the dominant group, and selecting, using a demand threshold selector, the mean demand threshold as the monthly demand threshold; and

providing continuous battery-optimized demand charge minimized EMS operations based on the monthly demand threshold using a real-time controller configured for overriding the optimal charging/discharging profiles when a monthly demand threshold violation is detected.

10. The system as recited in claim 9 , wherein the clustering is executed using k-means unsupervised learning.

11. The system as recited in claim 9 , wherein the processor is further configured for performing, using the demand threshold optimizer, monthly demand threshold optimization based on monthly demand charge, average battery state-of-charge (SoC), and accumulated battery depth-of-discharge (DoD).

12. The system as recited in claim 9 , wherein the processor is further configured for controlling daily battery utilization, charging actions, and discharging actions as defined by the monthly demand threshold (DT), power demand by a consumer, and battery limits.

13. The system as recited in claim 9 , wherein the plurality of groups from the clustering include distinct groups for weekends, weekdays, and events.

14. The system as recited in claim 9 , wherein data utilized for the daily optimizations includes one or more of maximum power demand, average power demand, daily demand threshold, demand charge saving, battery discharged energy, and average battery state-of-charge (SoC).

15. The system as recited in claim 9 , wherein the optimal monthly demand charge threshold selection is based only on a received customer load profile, the load profile including features of maximum power demand and average power demand.

16. The system as recited in claim 9 , wherein the optimal monthly demand charge threshold selection is based on a received customer load profile and savings, the load profile and savings including features of maximum power demand, average power demand, demand threshold, demand charge saving, average battery state-of-charge (SoC), and accumulated discharged energy.

17. A non-transitory computer readable storage medium comprising a computer readable program for controlling behind-the meter energy storage/management systems (EMSs) for battery-optimized demand charge minimized operations, comprising:

determining an optimal monthly demand charge threshold based on at least one of a received customer load profile and a customer load profile and savings, the determining further comprising:

iteratively performing daily optimizations to determine a daily optimal demand threshold for each day of a month;

selecting a monthly demand threshold by clustering the daily optimal demand thresholds for each day of the month into one of a plurality of groups, and determining a dominant group representative of a load pattern for a next month; and

determining a mean demand threshold for the dominant group, and selecting the mean demand threshold as the monthly demand threshold; and

providing continuous battery-optimized demand charge minimized EMS operations based on the monthly demand threshold using a real-time controller configured for overriding the optimal charging/discharging profiles when a monthly demand threshold violation is detected.

18. The computer readable storage medium as recited in claim 17 , wherein the clustering is executed using k-means unsupervised learning.

19. The computer readable storage medium as recited in claim 17 , further comprising performing monthly demand threshold optimization based on monthly demand charge, average battery state-of-charge (SoC), and accumulated battery depth-of-discharge (DoD).

20. The computer readable storage medium as recited in claim 17 , further comprising modeling and controlling daily battery utilization, charging actions, and discharging actions as defined by the monthly demand threshold (DT), power demand by a consumer, and battery limits.