The task of the NILM is to perform decomposition of P (t) into appliance specific power signals in order to achieve disaggregated energy sensing. Electrical loads exhibits a unique energy consumption pattern often termed as ��load or appliance signatures��, that enables the disaggregation algorithms to discern and recognize appliance operations from the aggregated load measurements. Appliance identification is highly dependent on load signatures, which are further characterized by the appliance category. As proposed by [6], consumer appliances can be categorized based on their operational states as follows:Type-I: These are the appliances with only two states of operation (ON/OFF). Examples of such devices includes table lamp, toaster, etc.
Type-II: These are multi-state appliances with a finite number of operating states also referred to as Finite State Machines (FSM). Consumer appliances belonging to this category includes washing machine, stove burner etc. The switching pattern of these appliances is also repeatable, which makes it easier for the disaggregation algorithm to identify their operation.Type-III: The appliances belonging to this category are also known as Continuously Variable Devices (CVD) because of their variable power draw characteristics with no fixed number of states. The power drill and dimmer lights are examples of CVD��s with no repeatability in their power draw characteristics.
Hence it is very challenging for the NILM methods to disaggregate these type of appliance from the aggregated load measurements.
Type-IV: In [5,7] authors have highlighted another category of appliances that remain active throughout weeks or days Batimastat consuming energy at a constant rate and therefore referred to as ��permenant consumer devices��. Appliances such as hardwired smoke detector, telephone sets, cable TV receivers are amongst the devices belonging to this category.Figure 1.(a) General Entinostat framework of NILM approach (b) An aggregated load data obtained using single point of measurement; (c) Different load types based on their energy consumption pattern.The energy consumption pattern of different type of loads have been shown in Figure 1(c), which is further translated as an appliance feature to distinguish between different appliance categories. Research to date has tended to focus on defining load signatures tailored to the appliance categories listed above in order to characterize them in a best possible way for identification.