Method of utility usage divergence identification between recurring consumption data indicative of repeating periodic consumption patterns and utility consumption data

1. A computer-implemented method of analysing utility consumption at a utility supply location, the method performed at an analysis server and comprising the steps of:

receiving from a utility meter utility consumption data, the utility consumption data corresponding to utility consumption at the utility supply location over a time period to be analysed;

generating recurring consumption data indicative of periodic consumption patterns in the utility consumption data;

identifying divergences between the utility consumption data and the recurring consumption data;

computing a diagnostic measure indicative of irregular consumption based on the identified divergences; and

sending a control message to adjust, based on the diagnostic measure, one or both of: the utility supply from one or more suppliers; and the utility consumption of one or more utility supply locations;

wherein the control message includes one of:

a control command to utility supply or generation equipment to increase or decrease production or supply; and

a control command to control operation of utility consuming equipment;

wherein generating recurring consumption data comprises:

performing a frequency domain transform on the utility consumption data to compute a frequency domain representation of the consumption data;

discarding frequencies from the frequency domain representation according to a frequency discard criterion; and

performing an inverse frequency domain transform on the frequency domain representation comprising the non-discarded frequencies to create a time domain representation of the consumption data based on the non-discarded frequencies.

2. A method according to claim 1 , wherein generating recurring consumption data comprises performing a frequency analysis to determine frequency components of the utility consumption data.

3. A method according to claim 2 , comprising selecting one or more frequency components and generating the recurring consumption data based on the selected frequency components.

4. A method according to claim 1 , further comprising:

calculating a difference between the time domain representation of the consumption data based on the non-discarded frequencies and the received utility consumption data;

determining whether the calculated difference satisfies a correlation criterion; and if the calculated difference satisfies the correlation criterion, using the time domain representation of the consumption data based on the non-discarded frequencies as the recurring consumption data.

5. A method according to claim 4 , wherein determining whether the calculated difference satisfies a correlation criterion comprises:

calculating an autocorrelation function on the difference between the received utility consumption data and the time domain representation of the consumption data based on the non-discarded frequencies; and

determining whether the autocorrelation function satisfies the correlation criterion.

6. A method according to claim 5 , wherein determining whether the autocorrelation function satisfies the correlation criterion comprises:

identifying statistically significant peaks in the autocorrelation function;

computing a proportion of statistically significant peaks in the autocorrelation function; and

determining whether the proportion of statistically significant peaks is below a predetermined correlation threshold.

7. A method according to claim 1 , wherein the recurring consumption data comprises a times series of utility consumption data representative of repeating consumption patterns.

8. A method according to claim 1 , wherein identifying divergences between the utility consumption data and the recurring consumption data comprises:

determining a set of residual consumption data from the difference between the utility consumption data and the recurring consumption data; and

detecting signal features in the residual consumption data indicative of divergences.

9. A method according to claim 8 , wherein detecting signal features indicative of divergences comprises:

identifying spikes in the residual consumption data by:

identifying rising and/or falling edges in the residual consumption data.

10. A method according to claim 1 , wherein computing a diagnostic measure comprises:

determining a measure of the incidence of divergences between the utility consumption data and the recurring consumption data, wherein determining the incidence of divergences comprises:

dividing the time period into a plurality of time slots for analysis;

identifying time slots having at least a threshold number of divergences; and

determining the number or proportion of time slots having at least the threshold number of divergences.

11. A method according to claim 1 , wherein computing a diagnostic measure further comprises:

calculating the diagnostic measure as, or based on, a ratio of time slots having at least a threshold number of divergences to a total number of time slots in the time period.

12. A method according to claim 1 , further comprising the step of:

pre-processing the received utility consumption data to compensate for abnormalities in the received consumption data, and using the pre-processed utility consumption data when generating the recurring consumption data, identifying divergences between the utility consumption data and the recurring consumption data and/or computing the diagnostic measure.

13. A method according to claim 1 , comprising selecting one or more utility supply locations based on their respective diagnostic measures, the selected utility supply locations having diagnostic measures exceeding a threshold, and transmitting consumption control information to one or more of: the identified utility supply locations, users associated with the identified utility supply locations, and user devices associated with the identified utility supply locations.

14. A method according to claim 1 , further comprising the step of:

controlling a supply of a utility in a utility supply grid that supplies the utility supply location based on the diagnostic measure or based on diagnostic measures computed for a plurality of supply locations.

15. A method of controlling utility supply in a utility supply grid comprising:

analysing utility consumption using a method as set out in claim 1 for a plurality of supply locations to output a plurality of diagnostic measures;

creating or altering a supply grid control model based on the plurality of diagnostic measures;

sending one or more commands to supply grid equipment according to the control model.

16. A method according to claim 15 , wherein sending commands to supply grid equipment comprises sending commands to perform at least one of: change the capacity of the supply grid, and change the operation of energy generation equipment.

17. A method according to claim 1 , wherein the received utility consumption data is received via a communications network, the utility meter located at the utility supply location, and comprises consumption data values measured by the utility meter.

18. A computer system or computing device having a processor with associated memory for performing the method of:

receiving from a utility meter utility consumption data, the utility consumption data corresponding to utility consumption at a utility supply location over a time period to be analysed;

generating recurring consumption data indicative of periodic consumption patterns in the utility consumption data;

identifying divergences between the utility consumption data and the recurring consumption data;

computing a diagnostic measure indicative of irregular consumption based on the identified divergences; and

sending a control message to adjust, based on the diagnostic measure, one or both of: the utility supply from one or more suppliers; and the utility consumption of one or more utility supply locations;

wherein the control message includes one of:

a control command to utility supply or generation equipment to increase or decrease production or supply; and

a control command to control operation of utility consuming equipment;

wherein generating recurring consumption data comprises:

performing a frequency domain transform on the utility consumption data to compute a frequency domain representation of the consumption data;

discarding frequencies from the frequency domain representation according to a frequency discard criterion; and

performing an inverse frequency domain transform on the frequency domain representation comprising the non-discarded frequencies to create a time domain representation of the consumption data based on the non-discarded frequencies.

19. A non-transitory computer readable medium comprising software code adapted, when executed on a data processing apparatus, to perform the method of:

receiving from a utility meter utility consumption data, the utility consumption data corresponding to utility consumption at a utility supply location over a time period to be analysed;

generating recurring consumption data indicative of periodic consumption patterns in the utility consumption data;

identifying divergences between the utility consumption data and the recurring consumption data;

computing a diagnostic measure indicative of irregular consumption based on the identified divergences; and

sending a control message to adjust, based on the diagnostic measure, one or both of: the utility supply from one or more suppliers; and the utility consumption of one or more utility supply locations;

wherein the control message includes one of:

a control command to utility supply or generation equipment to increase or decrease production or supply; and

a control command to control operation of utility consuming equipment;

wherein generating recurring consumption data comprises:

performing a frequency domain transform on the utility consumption data to compute a frequency domain representation of the consumption data;

discarding frequencies from the frequency domain representation according to a frequency discard criterion; and

performing an inverse frequency domain transform on the frequency domain representation comprising the non-discarded frequencies to create a time domain representation of the consumption data based on the non-discarded frequencies.