Microgrids and Virtual Power Plants

One of the key challenges for the future power grid will be coping with Distributed Energy Resources (DERs) such as Combined Heat and Power (CHP), photovoltaic systems, and wind turbines. In order to address this challenge, new concepts are currently being developed. However, the market is still in its early stages, and as yet there is no commonly accepted terminology. The terms “microgrid,” “Virtual Power Plant” (VPP), “embedded generation,” and “smart distribution network” all describe similar concepts. What these concepts have in common is that they all incorporate the idea of both energy resources and loads. Most VPP concepts also include the notion of load management, even though the term “power plant” generally refers to energy sources only.

The U.S. Department of Energy Microgrid Exchange Group defines a microgrid as “a group of interconnected loads and distributed energy resources within clearly defined electrical boundaries that acts as a single controllable entity with respect to the grid. A microgrid can connect and disconnect from the grid to enable it to operate in both grid-connected or island mode.” [US1] 

The Microgrids Group at Berkeley Lab describes three key features of a microgrid [LBL1]:

• Design around total system energy requirements: Microgrids optimize the overall energy system of the end user. For example, Combined Heat and Power (CHP) systems can be used to limit heat waste by means of local thermal generation of electricity.
• Provision of a heterogeneous level of Power Quality and Reliability (PQR) to end users: Microgrids provide high power quality to critical loads (such as lighting), while less critical loads (like refrigeration or ventilation) can receive lower PQR depending on availability. This is different to macrogrids, where consistent service quality is an important goal.
• Presentation to the macrogrid as a single controlled entity: A microgrid presents itself to the surrounding distribution grid as a single controlled system, acting as a “model citizen” that helps to reduce congestion, offset the need for new generation, supply local voltage support, and respond to rapid changes in load levels.

All of these features of a microgrid implicitly assume the existence of a local control function that is independent of the macrogrid. For lack of a widely established term (and in order to explicitly include the aspect of load management), we will use the term “Virtual Power Plant/Microgrid Management System” (VPP/MMS) in the following paragraphs. This assumes that microgrids can be hierarchical and that a single VPP can manage multiple microgrids.

VPP/MMS: Functional Overview

The key functionalities of a Virtual Power Plant/Microgrid Management System (VPP/MMS) include integration of different, heterogeneous energy sources and loads (“assets”), asset data management, energy management, and integration with external partners. Using our Asset Integration Architecture (AIA) once again as a reference, the following diagram provides an overview of a typical VPP/MMS.

Integration with the different energy sources and loads (“assets,” in the Ignite | IoT terminology) can either occur through the installation of a local gateway and/or agent directly on the asset, or – if the asset already provides remote integration capabilities – through a connector in the backend.

The VPP/MMS backend typically provides some kind of asset management functionality, which enables configuration and administration of the various integrated assets. Asset-related data, such as master data and asset history including events, faults, and time series data, are managed centrally.

Building on this asset data, the VPP/MMS implements the logic required to manage the different assets and their energy supply and consumption levels. This includes modeling and forecasting, scheduling, and real-time optimization. As most microgrids will not be able to function completely autonomously, the VPP/MMS must also be integrated with external systems to receive information such as weather forecast data and market prices. In addition, the VPP/MMS will in many cases be integrated with external processes such as energy trading, billing, and other processes related to TSOs/DSOs (Transmission and Distribution System Operators).

Case Study: Smart City Rheintal

The Vorarlberg Rhine Valley (Rheintal) is a densely populated metropolitan area in Austria, with a population of approximately 240,000. The Smart City Rheintal project was initiated by the federal state of Vorarlberg and the largest local utility, illwerke vkw. The goal is to leverage local renewable resources (such as hydropower) in order to make the region as energy-independent as possible by 2050 [SCR14].

One major subproject within this initiative is tasked with building a Virtual Power Plant (VPP) to help balance out the different energy sources and loads involved. The current focus is more on functional integration rather than scalability. To date, the project has integrated the following components:

• Photovoltaic systems (PV)
• Consumer devices, including heat pumps and electric boilers equipped with storage units
• Electromobility infrastructure, including a charging infrastructure and electric vehicles from car-sharing companies
• Battery storage systems

Forecasting Photovoltaic Energy

The project is currently integrating PV facilities from one district into the VPP, with others soon to follow. Aggregated data is used to calculate day-ahead forecasts for PV power generation. The VPP monitors the facility’s actual power output every 15 minutes. This is helpful for improving the accuracy of the forecasts. The figure below shows a comparison of the forecast (red) and actual (blue) power output of a PV facility over a period of several days.

Consumer Devices: Load Management

The project has selected a number of flexible consumer devices that support active load management. These devices are less critical and support lower Power Quality and Reliability (PQR – see discussion above). Moreover, they also support external control interfaces, which is a prerequisite for integrating them into an active load management scenario.

The basic idea is to use these more flexible consumer devices to take advantage of peak energy situations and then power down when supply is low. In order to support this, the VPP creates an operations schedule using the generation forecast series for the next 24 hours. The VPP coordinates technical integration, integration of external data (for instance from the EXAA electricity exchange in Vienna), definition and implementation of threshold values, and prioritization of consumers.

Electromobility Infrastructure

For a smart grid, integration with the electromobility infrastructure is interesting for two different reasons:

• Electric vehicles (EVs) as power consumers: During the charging process, EVs act as power consumers (from the grid’s perspective).
• EVs as power storage units: The batteries in EVs can also be used as power storage units, with the limitation that these units are only accessible as long as an EV is connected to a charging station.

The initial focus of the Smart City Rheintal project was to integrate EVs as power consumers. To support this, the project has integrated several charging stations operated by car-sharing providers into the VPP. In order to generate a viable charging schedule, the VPP must consider various inputs, including local power-generation forecasts, car reservation data, EV power requirements and State-of-Charge (SOC) data, and the load profiles of the charging station.

The figure below provides an overview of the different elements of an EV infrastructure and how they integrate with each other in the context of a microgrid.

Lessons Learned

Michael Schlauch from Bosch Software Innovations is the Bosch project manager supporting the Smart City Rheintal project. Here are his two key takeaways from the project:

“We still see major room for improvement in the forecasting process. That applies to the generation forecasts for the PV facilities as well as the consumption forecasts for consumers at home. Here, longer time frames are needed that go beyond the end of the project being funded. That’s the only way to extrapolate practical and individualized consumption profiles for particular housing units, for example. Thanks to our collaboration with Vorarlberg University of Applied Sciences and specialized IT companies, we are well on our way to obtaining this type of forecast for a wide range of regional and structural parameters.”

“Consumer devices such as washing machines are too small to make any reasonable headway with demand side management. Devices integrated into the VPP have to be more flexible. This holds true for heat pumps as well as for electric boilers. And, of course, it’s also helpful when we can bundle together facilities with widely varying degrees of flexibility.”