Economic dispatch of thermoelectric systems using mathematical programming techniques.
The Brazilian Electric System (BES), typically hydrothermal, has an operational profile formulated by the basic strategy of economic optimization, seeking to determine which thermoelectric generating units and which quantities of energy each should produce, in order to supply the demand at minimum costs. The fuel cost of thermoelectric units in the problem of economic dispatch (ED), can be modeled through a quadratic function, however, due to the realistic effects observed in the operation of the units, mainly due to side effects, such as: the effect from the valve point, operations with multiple fuels and zones of prohibitive operations; impose a higher level of complexity on ED. The problem presents non-linear, non-convex, discontinuous and non-differentiable functions, therefore, it is essential to use reliable optimization tools, capable of providing speed and precision in obtaining the optimal values of the problem addressed. In this scenario, the present work aims to propose a classical optimization technique implemented from mathematical programming and based on Mixed Integer Linear Programming, using Branch-and-Bound algorithm, in AMPL programming language in addition to the Knitro and your tools for optimization. To this end, case studies were widely adopted in the specialized literature, where thermoelectric units were modeled, considering their operational restrictions, employing several renowned optimization techniques, comparing them with the proposed technique. The results of the comparative analysis for the solution of the ED problem, show that the technique has achieved considerable success in the optimization of the thermal systems analyzed, with results superior to or close to those obtained by the techniques covered in the comparative study, and with satisfactory processing times.