Distributed Economic Dispatch with Valve-Point Effect via LPV-based Modified Incremental Cost Consensus Approach

This research paper investigates a new distributed consensus-oriented optimization approach to minimize the fuel cost of distributed generation facilities subject to complex nonlin-ear valve-point effect. Valve-point effect occurs due to sequential opening of fuel control-valves, which causes ripples in the smooth quadratic cost function. To respond to this nonlinear optimization problem, a leaderless distributed consensus protocol is designed using modified incremental cost as a consensus variable along with linear parameter varying (LPV) modelling. Unlike existing techniques, the proposed approach is based on a multi-agent system framework, which appreciably reduces the communication and computation burden. The involved mathematical analysis unfolds that the optimal solution is attainable, when a consensus is achieved for modified incremental costs of all generation units and a constraint of balancing the supply and demand is validated. The convergence of modified incremental costs to a common value is illustrated using Lyapunov stability theorem and properties of algebraic graph theory. Simulations are carried out on a ten-unit benchmark test system, and the results show that the proposed approach is able to deal with the distributed economic dispatch subject to valve-point effect.