Method and system employing graphical electric load categorization to identify one of a plurality of different electric load types

1. A system for a plurality of different electric loads, the system comprising:

a plurality of sensors structured to sense a voltage signal and a current signal for each of the different electric loads;

a hierarchical load feature database comprising a plurality of layers, with one of said layers including a plurality of different load categories; and

a processor structured to:

acquire a voltage waveform and a current waveform from the sensors for a corresponding one of the different electric loads;

map a voltage-current trajectory to a grid including a plurality of cells, each of said cells being assigned a binary value of zero or one based on whether the voltage-current trajectory passes through each of said cells;

extract a plurality of different features from the mapped grid of cells as a graphical signature of the corresponding one of the different electric loads;

derive a category of the corresponding one of the different electric loads from the hierarchical load feature database; and

identify one of a plurality of different electric load types for the corresponding one of the different electric loads.

2. The system of claim 1 wherein the hierarchical load feature database comprises three of said layers; wherein a first layer of said layers includes the different load categories; wherein a second layer of said layers includes a plurality of different load sub-categories for each of said different load categories; and wherein a third layer of said layers includes the different electric load types for said different load sub-categories.

3. The system of claim 1 wherein the different load categories include resistive loads, reactive predominant loads, electronic loads with a power factor correction circuit, electronic loads without a power factor correction circuit, electric loads including a linear power supply using a transformer to boost voltage, phase angle controllable loads, and complex structures.

4. The system of claim 1 wherein said different features are eight different features.

5. The system of claim 1 wherein said grid includes a first horizontal axis defining a count of said cells and a second vertical axis defining said count of said cells; wherein said count is a positive plural integer N; wherein N=2Δ+1; and wherein said different features are selected from the group consisting of: (1) a binary value of one of the cells at a first location of the first horizontal axis and a location Δ+1 of the second vertical axis; (2) a binary value of one of the cells at a location Δ+1 of the first horizontal axis and the location Δ+1 of the second vertical axis; (3) a binary value determined by multiplication of binary values of all cells along a diagonal line in the grid from a lower left corner to an upper right corner to indicate whether the voltage-current trajectory is linear or non-linear; (4) a number of continuums of said cells with adjacent coordinates having a binary value of one within all of said cells to indicate a number of intersections of the voltage-current trajectory and an average of a maximal value and a minimal value of the voltage waveform corresponding to the first horizontal axis; (5) a count of zero or a number of self-crossing intersections of the voltage-current trajectory; (6) a number of intersections of the voltage-current trajectory with 1.3 times the average of the maximal value and the minimal value of the voltage waveform corresponding to the first horizontal axis; (7) existence of a central horizontal line-segment that occupies at least 30% of the entire first horizontal axis and if at least 50% of the central horizontal line-segment overlaps with part of the voltage-current trajectory; and (8) a binary value of one of the cells at the location Δ+1 of the first horizontal axis and a first location of the second vertical axis.

6. The system of claim 1 wherein the voltage waveform and the current waveform each include a total of K data points of the form (v k ,i k ), wherein: k=1, . . . , K; wherein v k and i k are a voltage value and a current value of a sample data point k, respectively; wherein maximal and minimal values of the voltage waveform and the current waveform are computed from:

v max =max v k ,

v min =min v k ,

i max =max i k ,

i min =min i k ,

v

0

=

1

2

⁢

(

v

max

+

v

min

)

,

and

i

0

=

1

2

⁢

(

i

max

+

i

min

)

;

wherein v 0 and i 0 are both averages of the corresponding maximal and minimal values, which form central points of the grid of said cells; wherein Δ defines size of said grid; wherein

ⅆ

v

=

v

max

-

v

0

Δ

ⅆ

i

=

i

max

-

i

0

Δ

;

wherein said processor is further structured to generate two sequences from the voltage waveform and the current waveform as:

{ v 0 −dv·Δ,v 0 −dv ·(Δ−1), . . . , v 0 , . . . ,v 0 +dv ·(Δ−1), v 0 +dv ·Δ}, and

{ i 0 −di·Δ,i 0 −di ·(Δ−1), . . . , i 0 , . . . ,i 0 +di ·(Δ−1), i 0 +di·Δ};

wherein each of said two sequences has N=2Δ+1 elements; wherein the grid of said cells includes a first axis having N of said cells and a second axis having N of said cells; and wherein each of said cells is assigned a positional value (v 0 +dv·x, i 0 +di·y) and a binary model value B x,y , which is initialized to be 0.

7. The system of claim 6 wherein said processor is further structured to map one-half cycle of the voltage waveform and the current waveform to the grid of said cells and to assign each of said K data points to a corresponding one of said cells with said binary model value B x,y of 1.

8. The system of claim 1 wherein said processor is further structured to determine a number of self-crossing intersections contained by the mapped voltage-current trajectory.

9. The system of claim 1 wherein the category of the corresponding one of the different electric load types is derived from a Supervised Self-Organizing Map.

10. The system of claim 1 wherein the category of the corresponding one of the different electric load types is derived from a self-organizing map or a self-organizing feature map trained using competitive learning.

11. A method of identifying load types for a plurality of different electric loads, said method comprising:

sensing a voltage signal and a current signal for each of the different electric loads;

providing a hierarchical load feature database comprising a plurality of layers, with one of said layers including a plurality of different load categories;

providing a processor;

acquiring, with the processor, a voltage waveform and a current waveform for a corresponding one of the different electric loads;

mapping, with the processor, a voltage-current trajectory to a grid including a plurality of cells, each of said cells being assigned a binary value of zero or one based on whether the voltage-current trajectory passes through each of said cells;

extracting, with the processor, a plurality of different features from the mapped grid of cells as a graphical signature of the corresponding one of the different electric loads;

deriving, with the processor, a category of the corresponding one of the different electric loads from the hierarchical load feature database; and

identifying, with the processor, one of a plurality of different electric load types for the corresponding one of the different electric loads.

12. The method of claim 11 further comprising:

employing three of said layers in said hierarchical load feature database;

including the different load categories in a first layer of said layers;

including a plurality of different load sub-categories for each of said different load categories in a second layer of said layers; and

including the different electric load types for said different load sub-categories in a third layer of said layers.

13. The method of claim 11 further comprising:

including resistive loads, reactive predominant loads, electronic loads with a power factor correction circuit, electronic loads without a power factor correction circuit, electric loads including a linear power supply using a transformer to boost voltage, phase angle controllable loads, and complex structures as the different load categories.

14. The method of claim 11 further comprising:

employing as said different features eight different features.

15. The method of claim 11 further comprising:

including a first horizontal axis of said grid defining a count of said cells and a second vertical axis defining said count of said cells;

employing a positive plural integer N as said count, with N=2Δ+1; and

selecting said different features from the group consisting of: (1) a binary value of one of the cells at a first location of the first horizontal axis and a location Δ+1 of the second vertical axis; (2) a binary value of one of the cells at a location Δ+1 of the first horizontal axis and the location Δ+1 of the second vertical axis; (3) a binary value determined by multiplication of binary values of all cells along a diagonal line in the grid from a lower left corner to an upper right corner to indicate whether the voltage-current trajectory is linear or non-linear; (4) a number of continuums of said cells with adjacent coordinates having a binary value of one within all of said cells to indicate a number of intersections of the voltage-current trajectory and an average of a maximal value and a minimal value of the voltage waveform corresponding to the first horizontal axis; (5) a count of zero or a number of self-crossing intersections of the voltage-current trajectory; (6) a number of intersections of the voltage-current trajectory with 1.3 times the average of the maximal value and the minimal value of the voltage waveform corresponding to the first horizontal axis; (7) existence of a central horizontal line-segment that occupies at least 30% of the entire first horizontal axis and if at least 50% of the central horizontal line-segment overlaps with part of the voltage-current trajectory; and (8) a binary value of one of the cells at the location Δ+1 of the first horizontal axis and a first location of the second vertical axis.

16. The method of claim 11 further comprising:

including a total of K data points of the form (v k ,i k ) with each of the voltage waveform and the current waveform, wherein: k=1, . . . , K; wherein v k and i k are a voltage value and a current value of a sample data point k, respectively; wherein maximal and minimal values of the voltage waveform and the current waveform are computed from:

v max =max v k ,

v min =min v k ,

i max =max i k ,

i min =min i k ,

v

0

=

1

2

⁢

(

v

max

+

v

min

)

,

and

i

0

=

1

2

⁢

(

i

max

+

i

min

)

;

wherein v 0 and i 0 are both averages of the corresponding maximal and minimal values, which form central points of the grid of said cells; wherein Δ defines size of said grid; wherein

ⅆ

v

=

v

max

-

v

0

Δ

ⅆ

i

=

i

max

-

i

0

Δ

;

and

generating two sequences from the voltage waveform and the current waveform as:

{ v 0 −dv·Δ,v 0 −dv ·(Δ−1), . . . , v 0 , . . . ,v 0 +dv ·(Δ−1), v 0 +dv ·Δ}, and

{ i 0 −di·Δ,i 0 −di ·(Δ−1), . . . , i 0 , . . . ,i 0 +di ·(Δ−1), i 0 +di·Δ};

wherein each of said two sequences has N=2Δ+1 elements; wherein the grid of said cells includes a first axis having N of said cells and a second axis having N of said cells; and wherein each of said cells is assigned a positional value (v 0 +dv·x, i 0 +di·y) and a binary model value B x,y , which is initialized to be 0.

17. The method of claim 16 further comprising:

mapping one-half cycle of the voltage waveform and the current waveform to the grid of said cells; and

assigning each of said K data points to a corresponding one of said cells with said binary model value B x,y of 1.

18. The method of claim 11 further comprising:

determining a number of self-crossing intersections contained by the mapped voltage-current trajectory.

19. The method of claim 11 further comprising:

deriving the category of the corresponding one of the different electric load types from a Supervised Self-Organizing Map.

20. The method of claim 11 further comprising:

deriving the category of the corresponding one of the different electric load types from a self-organizing map or a self-organizing feature map trained using competitive learning.