Iterative Adaptive Dynamic Programming Adviced Campus Scale Social Energy System Management

Demand side management (DSM) schemes are becoming more and more sophisticated in modern power systems, in which energy consumption should be properly planned and controlled, together with social elements such as consumer experience, environmental influences, and management policy and rules. In this paper, inspired by the concept of “social energy”, we incorporate user experience evaluation, real-time electricity price calculation and energy consumption prediction into a unified campus-scale energy management system, to investigate the interactions among energy and social elements. The University of Denver's campus grid is used as the test bench, and energy consumption data were collected from 6 on-campus buildings to conduct the numerical experiments. Specifically, to deal with the “optimization-over-optimization” problem introduced by real-time pricing, and to tackle the high complexity problem brought by dynamic building consumption patterns, we implement neural network based iterative adaptive dynamic programming algorithm in this complex system to solve for the best overall system social cost. Experiment results show that the proposed algorithm is able to provide reasonable energy management advice in a practical socio-technical energy system.