Local Load Optimization in Smart Grids with Bayesian Networks
| Authors | |
|---|---|
| Year of publication | 2016 |
| Type | Article in Proceedings |
| Conference | The 2016 IEEE International Conference on Systems, Man, and Cybernetics (SMC 2016) |
| MU Faculty or unit | |
| Citation | |
| Doi | http://dx.doi.org/10.1109/SMC.2016.7844862 |
| Field | Informatics |
| Keywords | Smart Grids; Smart Meters; Bayesian Networks; Ripple Control; Load Management |
| Description | One of the main goals of the power distribution utilities is to provide stable supply of the power load. The growing popularity of smart grids, i.e. power grids enhanced with modern ICT, opened new possibilities to make the grid more efficient, secure and reliable. However, by introducing new elements into the infrastructure, such as small-scale photovoltaic power plants, the management of power load is becoming more challenging. In the paper, we address the issue of prevalent load management methods, which often lack sufficient flexibility to match growing complexity of the grid. We have developed a local load management component as an alternative to the still widely used ripple control technology. Our component is capable of individual water heating control by utilising customized TOU tariff schedules. The component uses Bayesian network model to incorporate uncertainties caused by inconsistent or incomplete information collected from smart meters. Our solution was deployed by a major energy distribution company in the Czech Republic in a real smart grid infrastructure consisting of more than 300 consumption points. In the case study, we confirmed viability of our approach as well as pointed out potential challenges remaining to be solved. |
| Related projects: |