Estimating Demand Response Flexibility of Smart Home Appliances via NILM Algorithm

Smart home appliances (SHAs) participation in demand response (DR) can effectively reduce the peak-to-valley difference of the power curve. Identifying the load patterns of SHAs and quantifying their DR flexibility can help grid operators take wise control measures. This paper proposes an unsupervised non-intrusive load monitoring (NILM) algorithm based on load signatures and combination optimization to separate different SHA loads. The electricity price-based DR model optimizes the usage behavior of SHAs and quantifies flexibility by comparing their behavior before and after DR. The validity of the NILM algorithm is verified on the real data set. The flexibility of different SHAs is compared based on the separated load data.