Distributed resource management by using population dynamics : wastewater treatment application

Abstract

Industrial processes commonly involve different sub-systems, each one of them with an associated controller. Moreover, although these sub-systems might be independent, the energy that the controllers use can be obtained from a limited common source. Furthermore, the total energy used in a control process might be limited because of economical reasons. A proper management of the energy resource would be achieved by having a central decision maker for this task using information of all the control signals. However, this centralized scheme would require an expensive communication network. As an alternative, the total available energy can be divided into the controllers within the process ensuring an upper bound of energy consumption. This work shows that dividing the resource throughout the controllers is not efficient regarding the control performance. Then, a distributed strategy is proposed to manage the available energy resource by using evolutionary game theory. Finally, the proposed strategy is implemented on a wastewater treatment system that considers a first stage given by an urban drainage system (UDS), and a final stage given by a non-drinkable water network (NDWN), i.e., water for streets cleaning, plants watering, etc. Simulation results show an improvement of the performance with the proposed methodology in comparison with an off-line resource management of energy.