In this line of research CEPEL develops methodologies and software programs related to:
(i) Application of distributed processing techniques – One of the biggest problems that are solved for real decision making in the electricity sector worldwide is for the Brazilian power planning, which uses the chain of energy simulation / optimization models developed by CEPEL. For this reason, since more than ten years the official versions of long and medium-term operation planning models have been using distributed processing techniques from the definition of parallelization strategies adopted in their formulations. This allows a drastic reduction in the models’ CPU times, making room for a continuous improvement in the mathematical modeling of the system, the time horizons and discretizations that are employed, the representation of uncertainties, and the solving strategies that are applied to solve the related problems.The parallelization strategies are performed through MPI (“Message Passing Interface”).
(ii) Development of models, algorithms and computer facilities for the integration of the energy sector software programs – it aims at integrating software models of the energy sector, mainly related to the models’ input and output data, thus assuring the consistency of data and studies. In addition, the graphical interface aims at facilitating the use of the models, mainly in relation to result analyses, study organization and control. Currently, it is based on the multi-platform technology, that allows local, local with parallel processing, remote with parallel processing, in remote clusters within the user’s company or in the cloud running , using the secure protocol SSH.
(iii) Support to the Intensive Computing Laboratory – To allow better use of the computer resources, Cepel develops user control methodologies and provides services and maintenance of hardware and software infrastructures necessary to those distributed processing techniques, while permanently improving the models.
Still in this line of research, CEPEL has been developing a new computational environment for the development and use of the optimization, simulation and statistical models developed within the DEA department, providing a unique database and allowing the different models to share the same input/output files. In this new environment all models shall be run on a Web interface.
Contact the responsible area via email: