A column generation method for optimal load management via control of electric water-heaters

Electric water heaters have been the focus of several previous studies where they are viewed as energy storing devices to be considered for peak load shaving within device interruption based load management programs. It is well known that such controls, when applied indiscriminately, can cause payback phenomena which could actually worsen existing peak loads. Thus, optimal switching strategies must be devised. A column generation based optimization scheme which combines the advantages of linear programming and dynamic programming approaches, both previously proposed in the load management literature, is presented. The method is successfully tested on a representative set of residential electric water heating load mixes. The results are analyzed in the paper, both from the computational load management point of view.