Latest Control Technology of Power Systems
The increasing integration of renewable power generation in power
systems poses new challenges to power system control. In this article, we describe a multi-layer power system architecture that addresses these challenges and comprises economic planning, supervisory power control, as well as voltage and frequency control as main layers. Several parts of this architecture have already been developed and tested on diverse components including building energy management, micro grid management and wind park control. We provide an overview of these applications and indicate white spots in this control architecture that require further research.
Need Control Technologies
Smart grids are one of the cornerstones for our future power system guaranteeing sustainability, cost efficiency and security of supply. They help to increase the share of power generation based on renewable sources thereby achieving sustainability. Smart grids also help to minimize necessary extension of the grid infrastructure resulting in minimal costs. The topic smart grids is very broad and has at least three technology pillars: control systems, power engineering, information and communication technology. All are equally important, but the control systems part is the crucial one, when thinking about power system stability and performance.
Power supply systems historically started as pure local, islanded infrastructures, e.g. for cities or industrial production sites. The interconnection of neighboring power systems has led to improved system reliability and economy of operation. These benefits have been recognized from the beginning, and thus interconnections have grown continuously. The results are very large systems of enormous complexity. For example, the synchronous grid of Continental Europe covers most countries of Europe as well as some countries in Africa and Asia.
To deal with these challenges, concepts for decentralized management of the decentral energy resources are discussed. But different to the situation in the early years of power technology, a cellular system architecture will still be connected with the outside world on the electrical and information side. The world is getting even more connected, since sustainability and cost efficiency will drive conversion technologies, like power-to-heat or electro-mobility, and result in a coordinated operation of our electricity, heat/cold or mobility infrastructure.
Multi-Layer Power System Architecture
The future multi-layer power system architecture developed by Siemens. The boxes describe entities like physical components, control and management systems, or market platforms. The coloring of the boxes indicates the different control tasks that are executed on these entities. The power system control architecture contains three major control levels:
— economic planning
— supervisory power control
— voltage and frequency control
These control levels extend the control tasks denoted today tertiary, secondary, and primary control. Moreover, the architecture contains the electric hardware and the associated low level controllers. This architecture is based on two fundamental principles: coordinated distributed control and self-similarity. For example, the energy market platform coordinates the distributed economic planning of generators and loads. Similarly, the power system control center coordinates the distributed power, voltage, and frequency control actions of the subordinate units. The self-similarity becomes obvious for the micro grid manager. It acts as one of multiple players in the distributed economic planning of the energy market platforms. At the same time, it coordinates the distributed power generation and consumption among its subordinate generators and loads. Among these generators and loads, there may be smart buildings that themselves have a building energy manager coordinating local generation and demand and offering the aggregation to the micro grid manager. These principles, coordinated distributed control and selfsimilarity, make this architecture particularly suitable for distributed generation in large-scale networks.
Economic Planning
The economic planning level has to achieve an economically efficient operation of the power system. Therefore, it has to match power supply and demand prediction, e.g. for the following 24 hours, of different units. The power supply units can be individual large power plants, wind farms, photovoltaic plants as well as groups of smaller power plants, e.g. aggregated by a micro grid manager. Today, power demand is largely uncontrolled; yet, we expect that demand side integration (DSI), e.g. for large buildings or industrial facilities, will play an important role to actively control power demand.
In large scale power systems, this economic optimization can be achieved by one or multiple energy market platforms similar to the European Energy Exchange (EEX) today. From a control perspective, this market-based economic optimization is a distributed optimization problem. A corresponding energy market platform allows individual units to optimize their local operation depending on incentives like market prices, e.g. using model-predictive control approaches.
Micro Grid and Island Grid Manager
Classical island grids use diesel generators for reliable power generation. Diesel hybrid power plants for electric island grids aim at replacing as much as possible expensive diesel fuel through cheaper renewable electricity from wind or PV. Such systems require at least a small amount of electricity storage, typically batteries, to guarantee short-term grid stability as well as to bridge the time required for diesel starts in the case of sudden renewable generation drops. The storage opens the possibility for optimizing plant schedules by shifting energy from times with surplus renewable generation to times with remaining residual load if the additional savings of fuel can compensate the cost of the storage system. This depends on prices for fuel and for storage. Moreover, the batteries allow for set point optimization within the diesel fleet.
— computing the trade-off between generation, storage, and consumption on the longer time scale, e.g. 24 hours; that is it coordinates the economic planning considering energy production, energy purchase, energy consumption, load and weather forecasts, and CO2 footprint;
— controlling voltage and current on the short time scale while satisfying as good as possible optimized schedules and negotiated contracts;
— estimating the grid status based on available measurements as input for voltage and current control;
— in on-grid mode: provide ancillary services for the main grid, e.g. primary reserve or reactive power; and
— in off-grid mode: maintaining the grid frequency
Operation Manager for Decentral Dispatch of Multi-Modal Energy Conversion Units
Another important field for economic planning are multi-modal energy conversion units like combined heat and power plants. The owners of these units are usually not willing to have their systems controlled by a central instance. As a result, the concept of an Operation Manager (OM) has been introduced at Siemens. The idea is that the central utility does not control a decentral energy conversion unit directly. The decentral units optimize their operation w.r.t. the economic objectives of its owner while taking the input signals from the central system operator into account. Thus, the incentive signals help to push the decentral unit into an operation mode favorable to the full power system, while the OM retains complete authority over the exact operation of the decentral unit. Operation managers have been developed for several decentral energy conversion systems, namely batteries and combined heat and power (CHP) units.
In case of a CHP unit, the incentive signal given to the unit is the predicted electricity price, which can be time varying. Additionally, the CHP OM receives the time course of the thermal load it has to provide as shown in Figure 3. Based on these signals, the OM generates an operation schedule that minimizes its overall costs
C(xCHP , xBH) = Cstart(xCHP ) + Cstart(xBH) + Cgas,CO2,OM(xCHP + xBH) − R(xCHP ).
A typical operation schedule of the CHP system both for the standard heat controlled mode and as calculated by the Operation Manager. It can be seen that the Operation Manager avoids a complete shutdown of the CHP unit, and the CHP operates preferably at times of high electricity price. It also requires only one startup of the backup heater. In a recent large-scale study, we have investigated the advantages of the CHP Operation Managers by running full-year simulations for three different European locations and their specific regulatory and meteorological conditions and several different CHP usage scenarios. Preliminary results show that in locations where CHP units are favorable at all (e.g. Germany with its feed-in tariff), our Operation Manager can save over 30% over standard heat controlled operation.
Building Energy Manager
Buildings consume a substantial fraction of energy at least in developed countries. Therefore, the main task of a building energy management system is the reduction of energy costs by selecting the cheapest energy source currently available and managing electric and thermal storage capacities while reducing losses. In addition, buildings can offer flexibility in their consumption which is basically a deviation from a previously agreed power schedule. Recent research focuses on model-predictive control (MPC) solutions applied to heating, ventilation and air condition (HVAC) control and meeting challenges in practical implementations. In the meantime, optimal self-consumption of local renewable electric power generation became also a topic of research and field trials. The main lever is the control of thermal and electric storage capacities thereby shifting a buildings net consumption or generation.
The main function blocks of a control solution as currently considered for building energy management solutions at Siemens. The function block “Optimization” provides an optimal control schedule for the building and its relevant controls, e.g. set points for electric heaters, (dis)charging of batteries, or heat pumps. The first part of this schedule is executed by a function block “realtime control” which takes safety and comfort margins as well as short-term fluctuations not considered in forecasts into account. The function block “Optimization” is executed again as soon as new forecasts are available or a significant deviation of observations from their prediction is detected. The input for this function block comes from system monitoring and from external forecasts.
Wind Park Control
Looking from the grid side, a wind farm can be seen as a single renewable energy generator unit which should deliver required active power and, if requested, support grid among others in voltage and frequency stabilization. The performance of the park is related to and measured at its coupling point to the grid, typically referred to as the Point of Common Coupling (PCC). The park control provides distributed power, voltage, and frequency control by
— delivering the required active power limited by the available wind power;
— participating in voltage, reactive power, and power factor regulation;
— enabling reactive power control without active power production; and
— supporting frequency control which requires power reserve.
On the other hand, looking from within the park, there is a collection of N generators, i.e. wind turbines, connected via the park’s own grid which might be of a non-negligible size. Hence, mathematically speaking, in a park there is a mapping from N individual wind turbines to the single point (PCC) where the overall park performance is measured. This N-to-1 mapping immediately indicates that there is a possible freedom in distributing the overall desired park active and reactive power production over N turbines. Figure below illustrates the typical hierarchical structure of the wind park control.
Experimental Results of a Building Energy Management System
A prototype of a comprehensive energy management system for commercial buildings has been developed within the project “Internet of Energy for Electric Mobility”. The system is designed to operate devices in a building, e.g. photovoltaic generators, batteries, or chargers for electric vehicles in a cost- or otherwise optimal way. To test this energy management solution, a lab demonstrator based on Siemens’ building automation system Desigo (Desigo Insight PC, Desigo PX controller plus various Desigo TX IO modules) has been set up. The so-called Smart Building Lab integrates PV inverters, a battery and battery charger, an HVAC system as well as various metering devices in order to emulate a typical commercial building. Figure 8 shows the schematic diagram of these components and their integration into the building automation system. In the future, the smart building lab and the smart grid laboratory will be integrated in order to emulate the interaction of a smart building with an innovative smart grid environment.
Power management of the complete set of electrical devices has been successfully tested in the project “Internet of Energy”. In a future project, SENSIBLE, the power management system will be extended to utilize the substantial source of flexibility offered by electro-thermal devices.
Conclusions and Open Issues
We have described a multi-layer power system architecture that addresses several challenges that result from the increasing integration of renewable power generation. Moreover, we have described the implementations of several components of this architecture. The results obtained so far are very promising and underline the practicability of this architecture.
Of course, many issues remain open for future work. To the authors’ opinion, future research should address easy to evaluate (analytical) stability criteria, decentralized (self-organizing) strategies for energy and power management, fault correcting control, stable and resilient operation of hybrid DC and AC power systems, and the efficient use of newly available sensor technology like phase measurement units (PMUs) for grid control. All of these topics do have special challenges when it comes to the domination of inverter driven in-feed, e.g. through photovoltaic or wind power.
The rise of power electronic equipment in our power grids can be seen as a major technological trend. Besides this, the coupling of our electrical power system to other energy sectors, like the heat/cold supply or the transportation infrastructure via electro-mobility, will have significant impact to the requirements for the control system.
References
Abdul Afram and Farrokh Janabi-Sharifi. Theory and applications of HVAC control systems–a review of model predictive control (MPC). Building and Environment, 72:343–355, 2014.
Joachim Balke. Integration of renewable energy in Europe. European Union, http://www.ee.washington.edu/research/pstca/, 2014.
Carsten Bose, Clemens Hoffmann, Claus Kern, and Michael Metzger. New principles of operating electrical distribution networks with a high degree of decentralized generation. In CIRED 2009. 20th International Conference and Exhibition on Electricity Distribution — Part 1, pages 1–4, June 2009.
Internet of Energy for Electric Mobility Project. http://www.artemis-ioe.eu/
Bundesnetzagentur. Monitoringbericht 2014. http://www.bundesnetzagentur.de, 2014.
IRENE Project. www.projekt-irene.de
Siemens AG. SICAM Microgrid Manager. http://w3.siemens.com/smartgrid/global/de/produkte-systemeloesungen/verteilnetzautomatisierung/microgrids/pages/sicam-microgrid-manager
Andrés Collazos, François Maréchal, and Conrad Gähler. Predictive optimal management method for the control of polygeneration systems. Computers & Chemical Engineering, 33(10):1584–1592, 2009.
F. Dörfler, M. R. Jovanović, M. Chertkov, and F. Bullo. Sparsity-promoting optimal wide-area control of power networks. IEEE Trans. Power Systems, 29(5):2281–2291, 2014.
Siemens Wind Turbine families. www.energy.siemens.com/hq/en/renewableenergy/wind-power/platforms, 2015
