Optimal Sensor Network Design to Monitor the Energy Performances of a Process Plant

Nowadays, there's an increased interest in industry for the energy audit, whose objective is identifying opportunities to improve energy efficiency. The analysis of the energy performances, for industrial processes, requires the knowledge of the physical quantities involved in the energy and mass balances, and therefore, the establishment of a sensor network. Many studies were conducted to solve this problem. However rare are those which helped reducing-the computing time substantially for medium and large size problems. This paper introduces a methodology for finding the optimal sensor network to be installed in a process plant. Using a parallelized combinatorial search associated with a Verification function, the proposed sequential methodology identifies a suboptimal set of possible sensor networks to be adopted, in ascending order of number of sensors. This methodology is tested on a starch concentrator (medium size problem), showing an effective reduction of computational time.