You are on page 1of 20

Paper presented at PowerGen Asia 27-29 Sep.

2011, Kuala Lumpur, Malaysia

Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid
Fazel Farahmand, Tanuj Khandelwal, JJ Dai, Farrokh Shokooh Operation Technology Inc., Irvine, California, USA

Abstract
It is inevitable that envision of power systems will be based on the Smart Grid. On the other hand the energy crisis enforces utilizing renewable and clean energy through small-scale Microgrids. Interactions of these two growing new technologies require development of more complicated control strategies. In another word while each Microgrid has to satisfy its own localized constraints and objectives, limitations and objectives enforced by the whole Smart Grid and/or Microgrids require extended and more conservative constraints. The first part of this paper provides a comprehensive review of Micro grids and Smart grids control strategies. This review covers clean energy stability analysis, generation control, generation forecasting, load forecasting, power management, load management, interchange analysis of Microgrids, multiobjective and multi-constraint optimization as well as load shedding on localized and global grids. At last, combinations of some of the discussed different control strategies through the application of an industrial tool on a real grid example is provided. Keywords: smart grid, microgrid, power generation management, load management.

PowerGen Asia 2011 - Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid.

Page 1

Paper presented at PowerGen Asia 27-29 Sep. 2011, Kuala Lumpur, Malaysia

1.

Introduction

According to the United States Department of Energy, today’s power system is 99.97 percent reliable. But just the 0.03 percent unreliability costs $150 billion a year. It is also a fact that if grids were just 5% more efficient, energy saving would be enough to permanently eliminate fuel and greenhouse gas emissions. That’s just enough to emphasize that increasing attention to reliability, efficiency, environmental issues and global competition has pushed forward all nations to smarter power systems. To take first steps in this way, every element of the power grid should act smart and be efficient. That’s the backbone of smart grid’s development, where its newly developed microgrid elements can be a zero-net energy commercial building, Plug-in (Hybrid) Electric Vehicles (PHEV), a wind farm or solar panels.

2.

Smart Grid Control Strategies

The aim of this paper is to develop an insight into the structure of micro and smart grids. Figure 1 provides major control strategies through an enterprise solution for power systems.

Figure 1: Platform of a typical Power System Management.
PowerGen Asia 2011 - Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid. Page 2

Figure 2 shows a flowchart that structures energy management. Kuala Lumpur. Figure 2: Flowchart of Smart Energy Management System. there are growing energy demands and increasing climate factors that have raised more interest for clean energy resources.Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid.1. Major grid requirements can be categorized but not limited to Interchange Transaction Scheduling. Automatic Generation Control and Economic Dispatching. Malaysia 2.1.1. distributed smart dispatching has to be network oriented. Economic Dispatching Economic Dispatching (ED) which these days is known as Smart Dispatching by smart grid users and also known as micro-dispatching by microgrid providers has the key role in smart grids. In other words energy providers must match demand and dispatch generation in the most economical way while system PowerGen Asia 2011 . 2011. 2. The major point of having economic dispatching is to maximize economical power generation and distribution benefits while maintaining customer control.Paper presented at PowerGen Asia 27-29 Sep. Compared to centralized control. Page 3 . These changes in the structure of the power industry have brought new challenges and concerns which require a trustable and smart energy dispatching tool. Energy Management System While power generation is distributed all across the grid. Any suggested energy management tool should be responsive to random or sudden demand variations.

Malaysia and customer’s requirements are considered. Objectives: Economic Dispatching objective function consists of all or some of the following items. 2011.          Minimize fuel costs Minimize transaction costs Minimize power (Active/Reactive) losses Minimize control movements/adjustments Minimize differences in system voltages Optimize voltage/line flow security indexes.Paper presented at PowerGen Asia 27-29 Sep. Figure 3: ED for minimization of fuel cost for three generation plants operated with different fuel costs and efficiency curves. Wide and interactive optimization objectives might need to be satisfied while system requirements and constraints must be taken into account (see Figure 3). .Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid. Adaptive solution and robust algorithms Minimize maintenance costs Minimize start-up costs Page 4 PowerGen Asia 2011 . Kuala Lumpur.

MVAR. MVA) o Transformer o Cable o Line o Reactor o Impedance 2. Automatic Generation Control For a stable grid. AVR. systems must assure balance between constantly varying load and generation. PowerGen Asia 2011 . AMP. MVAR.Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid. Kuala Lumpur.1. Malaysia Constraints: All or some of following constraints are enforced in ED calculations. PF) Bus voltage Interchange Transaction Schedules Branch flow (MW. 2011.Paper presented at PowerGen Asia 27-29 Sep. the frequency error at the interconnection has to be considered into account for the Area Control Error (ACE) calculation. While microgrid’s interchange commitment to the utility has to be satisfied. Page 5 .2.     Generation(MW.

Paper presented at PowerGen Asia 27-29 Sep. load variations tend to have non-zero ACE.3. ancillary services scheduling. Therefore interchange schedules are enforced to be agreed by connected microgrids.1. Kuala Lumpur. As it can be seen from Figure 4 dedicated for electricity power exchange and scheduling. and financial tracking of energy transactions. Figure 4: Schedules are stored in a database while generation is under effect of AGC modules and controlled based on meter readings. In such cases Area Control Error (ACE) Diversity Interchange (ADI) offers a means of reducing this control burden without sacrificing any participant in a group. Positive ACE requires decreasing generation while negative ACE enforces increasing generation. Interchange Transaction Scheduling For a smart grid environment the goal is to decrease the amount of uncertainties and to operate the system with the highest confidence. Page 6 . transaction management. Malaysia Ideally. PowerGen Asia 2011 . and energy cost analysis and reporting. ITS incorporates energy scheduling. Interchange Transaction Scheduling (ITS) provides the capability to schedule energy transfer from one control area to another while considering wheeling. 2.Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid. ACE should remain zero at all times. Normally momentary generation/load imbalances of each control area results in controlling error diversity. 2011.

Figure 5: Control of generators under interaction of system’s constraints enforced by ED and adjustments imposed by AGC. Proper distribution of consumption over time decreases the peak problem and consequently decreases the required size of the spinning reserve. ED and ITS would interact to adjust total generation of generators (See Figure 5). while their cost stays almost the same to maintain or operate. AGC. Each of these generators can be a microgrid subsystem distributed nationwide.2.Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid. Smart grid is not limited only to the generation and transmission. That’s because most of the time peaker plants will sit idle. 2011. Kuala Lumpur.Paper presented at PowerGen Asia 27-29 Sep. PowerGen Asia 2011 . 2. Smart users distribute their load consumption over time to have lower power peak and least cost impact from tariff enforced by the energy seller. Intelligent Load Management In order to distribute generation smoothly over time. That normally includes higher rates at hours of higher demand. different tariffs are enforced to users. Page 7 . Malaysia When all grid’s constraints and system’s objectives are defined.

Paper presented at PowerGen Asia 27-29 Sep. and the action should be fast. Malaysia Figure 6: Demand distribution over time for lower consumption peak. Breaker Interlock Scheme The Breaker Interlock Scheme is the simplest and fastest load shedding method to drop a constant amount of local loads in a Microgrid based on a specific action (See Figure 7). correct. 2011. An ideal load shedding scheme should be smart enough to act based on the latest system configuration. reliable and prioritized based with minimal possible shedding in minimal electrical distance. loss of generation and faults that end up with the generation deficiency condition. lightning strikes. Page 8 . a microsystem or a large portion of the smart grid might be islanded from the whole grid. For PowerGen Asia 2011 . Kuala Lumpur. Considering the fact that power systems normally operate in stable condition. Actions must be independent for every islanded subsystem. Mostly. load shedding is enforced in response to a system disturbance like switching errors. Major Load Shedding Methods: A.1.2. in cases of lightning strikes or switching surges. 2. Load Shedding By definition Load shedding is the amount of load that ideally must be shed from the grid in order to maintain stability and performance of the remaining active parts of the grid.Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid.

PowerGen Asia 2011 . where a grid might be affected by another grid this technique might not be enough or even applicable.Paper presented at PowerGen Asia 27-29 Sep. Figure 7: Breaker Interlock Load Shedding Scheme B. Frequency relays normally act in two stages. The first stage of action occurs right after a frequency drop is observed and the second stage occurs after a few cycles if system frequency didn’t recover. Under Frequency Relay (ANSI Device 81) Scheme One of the major effects of disturbance on a power system is frequency collapse or degradation. Figure 8: Under Frequency Load Shedding Scheme. Because of the objective nature of this method its response time is very slow. Page 9 . Frequency relays respond to this disturbance in a timely manner and shed loads. Kuala Lumpur. Malaysia larger power systems.Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid. 2011.

By controlling eigenvalues of the most sensitive bus. Intelligent Load Shedding (ILS) By having a complete and accurate knowledge of a system. the action of PLC is limited to the grid or Microgrid controlled by that specific PLC and for large smart grids. However. 2011.Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid. it still lacks system-wide operating conditions. E. ILS is based on a user-defined logic for system disturbances called triggers and prioritized loads to be shed (See Figure 9). Kuala Lumpur. Page 10 . However. Stability Based Load Shedding A load shedding algorithm based on voltage stability of the most sensitive bus is proposed. Its action depends on the written PLC logic and can shed an optimal amount of load. this method recovers a disturbed system. Figure 9: ETAP Intelligent Load Shedding Scheme.Paper presented at PowerGen Asia 27-29 Sep. Results demonstrate the PowerGen Asia 2011 . An industrial comparison of discussed methods is illustrated in Figure 10 comparing two factors of total final load that are shed and load-shedding response time. Malaysia C. D. Having correct and full knowledge of the whole grid’s configuration and real-time ratings is a must for convergence and performance of this method. Programmable Logic Controller-Based Load Shedding Programmable Logic Controller (PLC)-based load shedding by far is a much more reliable and faster solution than the two previous discussed methods. ETAP has proposed a new and intelligent load shedding (ILS) technique.

Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid. In order to achieve this goal. Kuala Lumpur.Paper presented at PowerGen Asia 27-29 Sep.2. The main criterion to run load restoration is the satisfaction of grid’s robust stability. Load Restoration After successful removal of a major disturbance. It’s in comparison to PLC based load shedding which cannot have an overlook on islanded systems. This confidence is obtained by running different studies e. 2. 2011. Page 11 . This action would minimize the duration of the outage while maintaining fault isolation without exceeding the capacities of alternate power sources and routes. Changing load priorities. transient stability PowerGen Asia 2011 . Minimize System Maintenance A key feature of Intelligent Load Shedding (ILS) is the ability to update and customize the load shedding logic without performing any PLC re-programming. faster and more optimal to shed load over conventional load shedding schemes. load flow analysis.2. and optimizing logic are some examples of actions that do not require reprogramming of the PLC. loads at different subsystems are restored step by step or in one action. adding or removing loads from the system. Trigger actions can end up to the islanding of a portion of the grid.g. Malaysia advantage of ILS for being much smarter. In that case ILS is intelligent enough to detect and act independently for the entire grid and every isolated microgrid. the maximum possible load should be restored automatically. Figure 10: Comparison of ILS vs other conventional load shedding techniques.

A good forecast has a direct and significant impact on costly generating unit startups and shutdowns. Malaysia analysis or eigenvalue analysis of the whole grid prior to the load acceptance at any microgrid or substation.Paper presented at PowerGen Asia 27-29 Sep. Kuala Lumpur. The restoration qualification feature evaluates the ability to restore loads by monitoring the system’s essential parameters. 2. Apart from forecasting of the amount of power that will be generated. This information helps to enable visibility of power consumption and increase the generation’s degree of control. PowerGen Asia 2011 .Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid. 2011. This makes the load forecasting an ideal tool for industrial users as well as utilities to reliably and accurately forecast future loading in the system. Page 12 . forecasting results can be used for plant sizing. Estimation and Forecasting While Measurement and Detection was a promising science a half century ago. estimation and prediction is tomorrow’s demand. Load Forecasting Utilization of Advanced Metering Infrastructure (AMI) techniques into smart grids has helped to accurately forecast loads in real-time.3.1. 2. Figure 11: Estimated pattern of load at a meter at a specific location of the grid. economic dispatching and operational purposes. managing system demand and scheduling system upgrades based on predicted load growth. energy purchases.3.

2. For example depending on the location of the wind farm. Page 13 . factors of level. Utilization of renewable energies like wind and solar impose additional variables that must be counted. regression methods or time series analysis have been introduced for short-term load forecasting. coefficient of variation and kurtosis. mean of generation. Figure 12: Wind rose which represents probability of direction and speed of the wind at a specific location of the grid.3. minimum. peak and even sparks of the wind energy are estimated. 2. the daily generation would change. Figure 12 shows one of the most commonly used statistical patterns. As it was discussed load forecasting was one of the major tools developed at the user level.Paper presented at PowerGen Asia 27-29 Sep. Generation Forecasting Many factors from generation to distribution and consumption should be considered to have a reliable forecast of power generation. Kuala Lumpur. PowerGen Asia 2011 . Already developed and in use in Germany.Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid. which is utilized for wind forecasting. Still the impact of new PHEV loads onto the grid is an ongoing research. skewness. average. 2011. Malaysia A great number of techniques like statistical methods (see Figure 11). There are two different approaches toward wind forecasting. Either future atmospheric variables are predicted and used toward calculations or statistical environmental data at a specific location is used to forecast generation.

2. Last but not least. These issues along with bad data and faulty meter readings have brought forward the idea to develop State and Load Estimation (SLE) for smart grid analysis rather than just having State Estimation. In addition. This makes wind forecasting methods applicable to solar forecasting. That’s why many forecasting service providers apply wind forecasting methods to solar generation forecasting. This fact makes wind generation and solar generation correlated phenomenon. PowerGen Asia 2011 . First.000 nodes. there is also a good chance to have faulty meters.000 to 100.3. There is a fundamental difference between traditional state estimation and distribution state estimation required for the smart grid. Kuala Lumpur. Figure 13: State and Load Estimation (SLE) uses limited number of meter readings. The basic problem of state estimation techniques was that the power system model should be known exactly. rather than a balanced power system. 2011. a smart grid has non-observable subsystems with many unmeasured load points. State and Load Estimation Early techniques of State Estimation (SE) dates back to the early 1970s. while topology of transmission systems was analyzed as a whole a smart grid has many independent substations and short or low impedance branches which raise numerical problems. Also rather than having a system with a few hundred or the maximum couple of thousand busses. Second. Another approach to energy forecasting is on error forecasting. Page 14 . winds are caused by a differential heating of the earth’s surface by the sun. smart grids ranges from 10.3.Paper presented at PowerGen Asia 27-29 Sep. bad data or measurement error at different points by the grid. which is used to determine how new clean sources of generation will fit into the market models. Malaysia Interestingly.Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid. grids are normally unbalanced.

two composite motors. twenty PTs. This power system model was designed in ETAP11 power system analysis and management software. Malaysia 3. connected to a power grid. Page 15 . with eight synchronous motors in different sizes and six lumped loads. Example An example of providing comprehensive multi-objective solutions to a system is shown in Figure 14. Figure 14: An example of 21-bus system with connected wind farms and solar panels. thirty two different kinds of meters (mulitmeter. PowerGen Asia 2011 .Paper presented at PowerGen Asia 27-29 Sep. 2011. Kuala Lumpur. four inverters. ten two-winding transformers. The example contains a power system with twenty one Buses. two synchronous generators. two three-winding transformer.Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid. twelve cables. four PV Arrays (See Figure16). two capacitors. 10 WECC Type one wind turbine generators (Details is shown in Figure 15) . ampmeter and voltmeter). Forty three HV circuit breakers. thirty CTs.

Kuala Lumpur. PowerGen Asia 2011 . 2011. Figure 16: An example of PV Array Editor. Malaysia Figure 15: Schematic of an individual WECC-Type 1 Wind Turbine Model in ETAP. Page 16 .Paper presented at PowerGen Asia 27-29 Sep.Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid.

2011. PowerGen Asia 2011 . light blue values are corresponding to real-time data. Malaysia For the sake of simplicity in displaying of One Line Diagram (OLD). As it can be seen in Figure 17. SLE values and running different real time modules in background.Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid. ETAP simulation software communicates with OPC servers to bring data from practically any provider of standard networking protocol or software platform and display values as a monitoring tool. Figure 17: The smart grid utilized with SCADA data. ETAP provides ability to design or put portion of model inside a composite network. Kuala Lumpur.Paper presented at PowerGen Asia 27-29 Sep. Page 17 .

Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid. For this example.Paper presented at PowerGen Asia 27-29 Sep. PowerGen Asia 2011 . In addition to above-mentioned control strategies.g. 2011. if any trigger action (e. A schematic of ILS editor is shown in Figure 19. minimal load would be shed at fastest experienced time. Figure 18: AGC parameters and ED objectives and constraints. Kuala Lumpur. Malaysia At the same time. State and load estimation is running in real-time on this model and its calculated values are shown in red. Figure19: Load shedding results in case of tripping a generator or utility. Utility trip) happens. Page 18 . schedules and rates. while AGC is set to have MW sharing for all generators (see Figure 18). power management is set to require minimal real power loss for economic dispatching. ETAP communicates with Microsoft® SQL Server to archive historian of data.

4. Power flows have always been available and the development of new information technologies over the web and synchrophasors have made data available over the globe. The larger the size of the network the more complicated are control strategies. constraints and interactive control strategies. the load is shed after 100 ms. Conclusion The smart grid is characterized as a unified power system with all of its connected substations and generation resources. However. so that a robust. 2011. for an unsupervised system. This complication makes the control of power system a chain of multivariable objectives. In such a case for the supervised grid.Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid. the grid will gradually become unstable and the system eventually runs into a full black out. PowerGen Asia 2011 . ILS sheds load and the system maintain its stability. Malaysia Transient stability analysis is also done to compare and validate two scenarios of trip without load shedding and trip with load shedding. It is essential that every facet and constraint in power system be taken into account. ILS suggests tripping CB12 and CB13. This makes the power transmission and distribution a road with two avenues of information and electricity flows. Kuala Lumpur. reliable and comprehensive solution is achieved. In this case as it can be seen from Figure 20. Page 19 . In both cases utility trip happens at the moment of one second. Figure 20: The left picture shows the bus voltage of major buses without load shedding.Paper presented at PowerGen Asia 27-29 Sep. The right picture shows the grid response in the presence of ILS. In case of load shedding.

to provide an enterprise solution to satisfy all interconnected and interacting requirements of every power grid. Its ability to provide load and generation estimation. archiving. Ultimately. Malaysia As a solution to all commitments and obligations of power systems. Page 20 . a distributed area of power supply and decentralized generation coming from clean resources will be the future of power systems. 2011. ETAP goes beyond offering just one or two tools.Paper presented at PowerGen Asia 27-29 Sep. Kuala Lumpur. forecasting.Interactive Objectives and Constraints Between Microgrid Elements of the Smart Grid. PowerGen Asia 2011 . real-time monitoring. simulation and optimization make it the brain for the grid to act smart.