This paper presents disaggregation of appliances' energy consumption for a residential house using Particle Swarm Optimization (PSO). The disaggregation process provides useful information about the time and duration of each appliance on when they are switched on, which can contribute to close monitoring of appliance activities in a house. Firstly, the data of aggregated electricity loads comprising six different appliances is measured and plotted based on total power (P total ) over time. The data is then formulated based on combinatorial optimization (CO) problem which represents the load disaggregation model that contains the sum power of all six appliances when they are switched on. The CO problem is solved by using PSO by setting up the population of particles to represent each appliance. The random number of each population is generated based on appliance's power state when switched on, where the total appliances power (P app ) is represented as the sum of all six particles data. Thus, by using the known value of appliance on state as optimization boundary parameters, the CO problem is evaluated by PSO following the Integral Squared Error (ISE) minimization problem of P total and P app . The optimization results give good convergence criteria with accurate percentage of time taken for every appliance in use.

This paper presents disaggregation of appliances’ energy consumption for a residential house using Particle Swarm Optimization (PSO). The disaggregation process provides useful information about the time and duration of each appliance on when they are switched on, which can contribute to close monitoring of appliance activities in a house. Firstly, the data of aggregated electricity loads comprising six different appliances is measured and plotted based on total power (Ptotal) over time. The data is then formulated based on combinatorial optimization (CO) problem which represents the load disaggregation model that contains the sum power of all six appliances when they are switched on. The CO problem is solved by using PSO by setting up the population of particles to represent each appliance. The random number of each population is generated based on appliance’s power state when switched on, where the total appliances power (Papp) is represented as the sum of all six particles data. Thus, by using the known value of appliance on state as optimization boundary parameters, the CO problem is evaluated by PSO following the Integral Squared Error (ISE) minimization problem of Ptotal and Papp. The optimization results give good convergence criteria with accurate percentage of time taken for every appliance in use. [206 words]