Methods and systems for determining power load disagrregation profile of a building

A system for determining a power load disaggregation profile of a building, the system comprising:

• a memory storing instructions;
• one or more Input/Output (I/O) interfaces;
• and
• one or more hardware processors coupled to the memory via the one or more I/O interfaces, wherein the one or more hardware processors are configured by the instructions to:
• receive sequentially, a plurality of data samples associated with a power consumption of the building having a plurality of appliance types measured using a smart meter installed in the building, and a time stamp for each of the plurality of data samples, wherein the building is connected to a power source or a power grid, through the smart meter, wherein each data sample of the plurality of data samples, is a power signal, obtained at a predefined low-sampling rate in minutes during the power consumption of the building;
• pre-process the plurality of data samples to obtain a plurality of pre-processed data samples, wherein the pre-processing includes filling missing data and outlier removal;
• detect and remove one or more background power load consumptions, from each of the plurality of pre-processed data samples, to obtain a plurality of resultant data samples;
• detect a plurality of positive events and a plurality of negative events, from amongst the plurality of resultant data samples, using the time stamp for each of the plurality of resultant data samples, based on a change in the power consumption;
• obtain one or more event pairs, from the plurality of positive events and the plurality of negative events, using an iterative event pairing technique;
• form one or more event pair clusters from the one or more event pairs, using a density-based clustering technique;
• and
• classify an appliance type of the plurality of appliance types, for each of the one or more event pair clusters, and estimating the power consumption for each of the classified appliance type, using a rule-based classification technique, to obtain the power load disaggregation profile of the building, wherein the rule-based classification technique comprises a set of rules and each of the set of rules is defined based on (i) appliance consumption levels of each appliance type in watts ranging from very low of less than predefined watts, low of more than the predefined watts, average, high, and extremely high, (ii) an usual operational time of each appliance type as only day, all day, only nights, (iii) an usage behaviour of each appliance type as low with 1 or 2 value, medium as greater than 2 and less than or equal to 4, and high with greater than and equal to 5, and (iv) seasonal appliance information of each appliance type, obtained using correlation analysis of annual power consumption of the building and local weather information, wherein loads are controlled based on the power load disaggregation profile using a predefined low-sampling rate for optimizing power usage.

One or more non-transitory machine-readable information storage mediums comprising one or more instructions which when executed by one or more hardware processors cause:

• receiving sequentially, a plurality of data samples associated with a power consumption of the building having a plurality of appliance types measured using a smart meter installed in the building, and a time stamp for each of the plurality of data samples, wherein the building is connected to a power source or a power grid, through the smart meter, wherein each data sample of the plurality of data samples, is a power signal, obtained at a predefined low-sampling rate in minutes during the power consumption of the building;
• pre-processing, the plurality of data samples to obtain a plurality of pre-processed data samples and the time stamp for each of the plurality of pre-processed data samples, wherein the pre-processing includes filling missing data and outlier removal;
• detecting and removing, one or more background power load consumptions, from each of the plurality of pre-processed data samples, to obtain a plurality of resultant data samples;
• detecting, a plurality of positive events and a plurality of negative events, from amongst the plurality of resultant data samples, using the time stamp for each of the plurality of resultant data samples, based on a change in the power consumption;
• obtaining, one or more event pairs, from the plurality of positive events and the plurality of negative events, using an iterative event pairing technique;
• forming, one or more event pair clusters from the one or more event pairs, using a density-based clustering technique;
• and
• classifying, an appliance type of the plurality of appliance types, for each of the one or more event pair clusters, and estimating the power consumption for each of the classified appliance type, using a rule-based classification technique, to obtain the power load disaggregation profile of the building, wherein the rule-based classification technique comprises a set of rules and each of the set of rules is defined based on (i) appliance consumption levels of each appliance type in watts ranging from very low of less than predefined watts, low of more than the predefined watts, average, high, and extremely high, (ii) an usual operational time of each appliance type as only day, all day, only nights, (iii) an usage behaviour of each appliance type as low with 1 or 2 value, medium as greater than 2 and less than or equal to 4, and high with greater than and equal to 5, and (iv) seasonal appliance information of each appliance type, obtained using correlation analysis of annual power consumption of the building and local weather information, wherein loads are controlled based on the power load disaggregation profile using a predefined low-sampling rate for optimizing power usage.