I linked two additional papers in our reports titled Nonlinear Power Flow Control Design of High Penetration Renewable Sources for AC Inverter Based Microgrids and A Predictive Engine for On-Line Optimal Microgrid Control. These papers extend the work described previously here and are the result of a continuing collaboration between Sandia, Michigan Tech, and myself at OptimoJoe.
In the first paper, we extend our previous work by generalizing our models from AC to DC microgrids. In fact, our current codes allow an arbitrary configuration of a combination of both DC and AC microgrid components. Part of the difficulty in this generalization is how to manage the interface between the AC and DC components while insuring that the resulting model remains a good candidate for a model predictive control. In our case, a direct-quadrature-zero (DQ0) transformation of the three-phase circuit combined with some clever transformations results in an optimization formulation that works well with Optizelle.
In our second paper, we discuss one possible method for predicting electrical loads. Recall, a model predictive or receding horizon control requires that we have some prediction of the future. Simply, if we intend to plan ahead, we need to know what we need to plan for. Often, we have a very good idea of what our future load demands will be. For example, if we turn on a piece of machinery and we know how much power this piece of machinery requires, then we can plan for the rise in demand. Other times, we don’t have this information, but we know the general trends to expect at this time in the day. Our paper discusses the latter and it seems to work well for this particular scenario.
We would like to thank both Sandia and MTU for their continued collaborations. Both groups are skilled and easy to work with and we highly recommend them. If any other groups are looking to solve some optimal control or design problems related to microgrids, we consult as well.