UC Berkeley

The advancement of sensing technologies for the power grid has allowed the development of new strategies for the control of distributed energy resources (DERs). In particular, the emergence of phasor measurement units (PMUs) designed for deployment at the distribution level has presented an exciting opportunity. These PMUs have enabled the development of phasor-based control (PBC), a strategy that formulates DER power dispatch in terms of voltage phasor targets to be tracked by local controllers.

This dissertation focuses on the optimal power flow (OPF) component of PBC’s supervisory control layer, which has previously been conceptualized and demonstrated in simulation on distribution networks. We expand its applicability to medium-voltage minigrids and microgrids operating in island mode, networks where PBC has the potential to deliver important benefits.

The work is carried out in two stages. After a discussion of PBC and other relevant background topics, we address one of the primary challenges to PBC at the medium voltage level: the need for extreme accuracy in the supervisory controller’s generation of phasor targets. This accuracy is achieved through an adaptation of an iterative OPF methodology that refines a linearized model of power flow through successive exchanges with a nonlinear solver. We discuss the changes that were made to both linear model and nonlinear solver, as well as the determination of phasor targets on networks that include tap-changing transformers and other realistic equipment. The accuracy of the adapted iterative method is then shown in simulation.

The second stage of the work covers the extension of our OPF implementation to islanded systems. We present a strategy for the treatment of the slack bus used by our nonlinear solver and apply it to several test cases in simulation. We then analyze a specific case in which our iterative solution method fails, and demonstrate the use of a penalty factor in our linearized OPF formulation as a means of overcoming that failure. A full, end-to-end implementation of PBC’s supervisory layer is then proposed and tested on a number of DER distributions at different feeder penetration levels.

We end with a presentation of data relevant to instrument-transformer-induced error in PMU measurements. This final portion stands alone from the primary work of the dissertation, but remains highly relevant to PBC. From an experimental deployment of two PMUs measuring an identical distribution-grid voltage, we determine a ratio of the errors induced by their individual potential transformers. Monitoring this quantity over the course of a year allows us to track the drift in those induced errors over time, motivating a discussion of the expected impact of error drift on PBC and the frequency with which transformers will need to be recalibrated in operational settings.