Mesa del Sol High-Penetration PV Testbed (Submitted to DOE SunShot program).

residential resources

We propose a project centered around a full-scale, highly-integrated testbed for research, training and product development in distribution-level high-penetration PV. Mesa del Sol, a sustainable planned community in Albuquerque, New Mexico, is the setting for several existing demonstration projects, publicly and privately funded, at a level of multiple 10s of $M, including utility-scale PV with battery storage, a state-of-the-art CERTS microgrid, building-scale PV / battery systems, all tied to a highly instrumented power distribution feeder. While the individual projects are currently used for successful research, they are not systematically related to one another to observe or control their interactions. We propose to use a hardware-in-loop (HIL) simulator to integrate these resources together, enabling the study of collective action of distributed resources. We further propose to combine a systems approach for control of distributed energy resources with a hands-on education and training program for aspiring and existing utility and energy industry professionals. We are partnering with Public Service Company of New Mexico, and with El Paso Electric.

Today, most utilities deal with the integration of PV-generated power on the grid in two ways: by ensuring that it is injected in what is essentially an infinite bus, or by curtailment. For example, in the Public Service Company of New Mexico (PNM) service territory, each application for PV interconnection is subject to analysis of its impact on the grid. If the additional PV system has the potential to disrupt power quality to unacceptable levels, then the interconnect may require costly upgrades. In locations where PV penetration is very high by design (e.g. southern Germany, where distributed PV generating capacity sometimes exceeds loads by a factor of 10, codes have been designed to mitigate the risk that devices could produce disturbances, first by a smooth power reduction according to a predefined curve, until generation is disconnected at 51.5 Hz. Neither approach is optimal: the former limits the penetration of PV to fairly low levels, while the latter wastes available clean energy.

It is often stated that, given sufficient storage, all these problems could be overcome. Unfortunately, storage is limited, expensive, and may not be able to fully compensate for renewable intermittency, if it does not have the right characteristics and is not distributed appropriately. Our project addresses the problem by utilizing existing and added storage strategically rather than ad hoc. First, we recognize that there are many types of storage other than batteries and flywheels, some of which may be hiding in plain sight, for example in the thermal mass of buildings or in controllable loads. Then, we apply an innovative control system design process developed by two of the co-PIs (Robinett and Wilson). The concepts developed result from the convergence of three research and development goals: to create a unifying metric to compare the value of different energy sources; to develop a new nonlinear control tool that applies power flow control, thermodynamics, and complex interconnected adaptive systems theory to the energy grid in a consistent way; and to apply theories developed to describe the collective motion of robots to a distributed, decentralized electric power grid. All three of these goals have important concepts in common: exergy flow, limit cycles, and balance between competing power flows.

This framework allows the process of control design to be viewed as a power flow control problem, balancing the power flowing into a system against that being dissipated within it and dependent on the power being stored in it, in an interplay between kinetic and potential energies. Its success has been shown in wind power, collective control of robots, and space systems. Control design techniques based on this framework were applied to DC microgrids with up to 100% PV penetration and AC grids with up to 100% wind penetration using swing equation models. The same framework will be extended to other aspects of high-renewable-penetration grids, including real-time markets. Using our approach, the path to technology development is decided by first setting the goals (e.g. 70% solar), then by determining the specifications of the storage needed by analysis, and finally categorizing the available storage and determining any additional needs.

By using modern tools for distance learning, the Mesa del Sol testbed can be made fully accessible to the entire consortium, and eventually to the wider community. The key issue revolves around maintaining a 'hands on' feel to the instruction. The difficulty of experiential learning in power systems is that electric power cannot be seen, only its generally unwanted effects (e.g. breakers tripping, fuses blowing). This is where the system described here excels. For example, we will be able to visualize power flow (in real and virtual systems) in real time, in slow motion, and in fast forward, as required by the application. For example, slow motion could be useful for demonstrating the operation of safety equipment (e.g. an inverter to a line fault), while fast forward mode could demonstrate the operation of real-time power retail markets. Moreover, the HIL simulation allows much greater flexibility than operating a hardware-only system, since the consequences of component or system failure are virtual. The concept is the same as that of a flight simulator for training airline pilots - extreme situations can be experienced without loss of safety. While there are many online academic offerings for power systems curricula, few offer a truly hands- on experience, and none are geared to provide training in emerging technologies. The curriculum proposed here will fill this gap, by providing a combination of theoretical knowledge and practical application. This training will not only apply to students entering the energy field but also to engineers and system operators already working in the field - allowing them to adapt to rapidly changing system characteristics