Disaggregation apparatus

1. A disaggregation apparatus for identifying an electrical consumer in an electrical network, wherein the electrical network comprises a power source and multiple electrical consumers, the disaggregation apparatus comprising:

an electrical signature providing unit for providing electrical signatures (i n ) of each of the electrical consumers, the electrical signature providing unit including a training method to switch on each electrical consumer one at the time and storing the corresponding electrical signature in the form of a current signature,

an electrical parameter determining unit for determining an overall electrical parameter (i tot,k ) of the electrical network, wherein the overall electrical current of the electrical network is measured,

an identification unit configured to normalize each of said electrical signatures (i n ), to generate a unity signature correlation matrix (R), which comprises correlations between each of said normalized electrical signatures with peak current equal to one of the respective electrical consumer's current, and identify an electrical consumer based at least partly upon the determined overall electrical parameter (i tot,k ) and the unity signature correlation matrix (R).

2. The disaggregation apparatus as defined in claim 1 , wherein the identification unit is configured to identify an electrical consumer by applying a multi user detection (MUD) method to the determined overall electrical parameter (i tot,k ) wherein the electrical consumers are regarded as users and the electrical network is regarded as communication channel.

3. The disaggregation apparatus as defined in claim 2 , wherein the identification unit is configured to identify an electrical consumer by applying a Zero Force (ZF) MUD method to the determined overall electrical parameter (i tot,k ).

4. The disaggregation apparatus as defined in claim 2 , wherein the identification unit is configured to identify an electrical consumer by applying a Minimum Mean Square Error (MMSE) MUD method to the determined overall electrical parameter (i tot,k ).

5. The disaggregation apparatus as defined in claim 1 , wherein the identification unit is configured to

determine an overall electrical parameter vector (i k ) by correlating the overall electrical parameter (i tot,k ) with each electrical signature (i n )

model the overall electrical parameter vector (i k ) as depending on a product of a signature correlation matrix (RA), which comprises correlations between the electrical signatures, and an activity vector (b k ) being indicative of the activity state of the respective electrical consumer,

determine the activity vector (b k ) such that a similarity measure, which is indicative of the degree of similarity between the modeled overall electrical parameter vector and the determined overall electrical parameter vector (i k ), is maximized,

identify an electrical consumer for which the determined activity vector (b k ) indicates an active state.

6. The disaggregation apparatus as defined in claim 5 , wherein the identification unit is configured to model the overall electrical parameter vector (i k ) as depending on a sum of i) the product of the signature correlation matrix (RA) and the activity vector (b k ), and ii) a noise vector (ση).

7. The disaggregation apparatus as defined in claim 1 , wherein the identification unit ( 9 ) is configured to

determine an overall electrical parameter vector (i k ) by correlating the overall electrical parameter (i tot,k ) with each electrical signature (i n ),

determine a modified overall electrical parameter vector (ĩ k ) depending on a difference between the overall electrical parameter vector (i k ) and a product of a signature correlation matrix (RA), which comprises correlations between the electrical signatures, and a unit vector ( 1 ),

model the modified overall electrical parameter vector (ĩ k ) as depending on a product of the signature correlation matrix (RA) and a modified activity vector ({tilde over (b)} k ), which depends on a difference between an activity vector (b k ), which is indicative of the activity state of the respective electrical consumer, and a constant,

determine the modified activity vector ({tilde over (b)} k ) such that a similarity measure, which is indicative of the degree of similarity between the modeled modified overall electrical parameter vector and the determined modified overall electrical parameter vector (ĩ k ), is maximized,

identify an electrical consumer for which the determined modified activity vector ({tilde over (b)} k ) indicates an active state.

8. The disaggregation apparatus as defined in claim 1 , wherein the identification unit is configured to determine an activity vector (b k ), which is indicative of the activity state of the respective electrical consumer, depending on the sign of a term comprising a product of an inverse unity signature correlation matrix (R −1 ), which comprises correlations between normalized electrical signatures (i n ), and a modified overall electrical parameter vector (ĩ k ) wherein the modified overall electrical parameter vector (ĩ k ) is determined depending on a difference between an overall electrical parameter vector (i k ) and a product of

a) the unity signature correlation matrix (R),

b) an amplitude matrix (A), which comprises amplitudes (A n ) of a part of the overall electrical parameter caused by the respective electrical consumer, and

c) a unit vector ( 1 ),

wherein the overall electrical parameter vector (i k ) is determined by correlating the overall electrical parameter (i tot,k ) with each normalized electrical signature (i n ).

9. The disaggregation apparatus as defined in claim 8 , wherein the identification unit is configured to model the modified overall electrical parameter vector (ĩ k ) as depending on a sum of i) the product of the unity signature correlation matrix (R), the amplitude matrix (A) and the activity vector (b k ), and ii) a noise vector (ση).

10. A disaggregation computer program for identifying an electrical consumer in an electrical network, wherein the electrical network comprises a power source and multiple electrical consumers, the disaggregation computer program comprising program code means for causing a disaggregation apparatus as defined in claim 1 to carry out the steps of the disaggregation method, when the computer program is run on a computer controlling the disaggregation apparatus.

11. A disaggregation method for identifying an electrical consumer in an electrical network, wherein the electrical network comprises a power source and multiple electrical consumers, the disaggregation method comprising:

providing electrical signatures of each of the electrical consumers, wherein the providing electrical signatures includes a training method that includes switching on each electrical consumer one at the time and storing the corresponding electrical signature in the form of the a current signature,

normalizing each of said electrical signatures,

generating a unity signature correlation matrix, which comprises correlations between each of said normalized electrical signatures with peak current equal to one of the respective electrical consumer's current,

determining an overall electrical parameter of the electrical network, by measuring the overall electrical current of the electrical network

identifying an electrical consumer depending on the determined overall electrical parameter and the unity signature correlation matrix.