A Data-driven Assessment Model for Demand Response Participation Benefit of Industries

Abstract

Demand Response (DR) is an incentivized program by the utility operator to provide an opportunity for consumers to play a significant role in the electric grid operation by shifting or reducing loads during peak periods. This work proposes a data-driven methodology that only uses smart meter data to identify load flexibility in industrial loads of consumer and cost-saving potential from participating in a DR program. The first step of the methodology involves an unsupervised clustering of historical demand loads data based on <tex>$K$</tex> -means algorithm to identify the energy usage behavior of an industrial consumer. An operation demand flexibility boundary is then calculated from the identified clusters. These boundaries are the flexible region where demand load ramp-up and ramp-down can be are achievable. Two DR participation scenarios (i.e., Passive and Active DR participation) based on Linear Constrained Optimization are designed where optimal daily electrical demand trajectory under DR participation scenario is estimated to evaluate the net benefit of DR participation. The case study of an electronics factory indicates that 4&#x0025; &#x2013; 7&#x0025; monthly net benefit can be achieved from passive DR participation, and 14&#x0025; &#x2013; 19&#x0025; monthly net benefit can be achieved from active DR participation. This methodology provides industrial consumers with a non-intrusive assessment of electrical load flexibility potential and associated DR participation benefit without going through the physical onsite audit process.