J Control Autom Electr Syst (2015) 26:96–104 DOI 10.1007/s40313-014-0158-y Development of a Smart Grid Simulation Environment, Part II: Implementation of the Advanced Distribution Management System Jônatas Boás Leite · José Roberto Sanches Mantovani Received: 26 September 2014 / Accepted: 30 September 2014 / Published online: 17 October 2014 © Brazilian Society for Automatics–SBA 2014 Abstract A smart grid is an electric power system with high levels of automation, dispersed generation, intelligent monitoring, and control; however, because the distribution networks in operation today do not have these characteris- tics, studies concerning the control, planning, and operation of the smart grid are difficult to perform. To overcome these practical difficulties, studies investigating the smart grid can be conducted in computational environments that are able to reproduce meticulously the electrical and communicational behaviors expected by the smart grid. Therefore, in this paper, the development of a platform to simulate the advanced dis- tribution management system (DMS) from a reference model for smart grid, which has seven layers, is proposed. Part II of this paper refers to the four upper layers. First, it describes the system layer architecture where all computational hardware is virtualized to create the smart grid simulation environ- ment. Subsequently, the three logical layers, Model, Analy- sis, and Intelligence, are considered in the construction of the advanced DMS supervisor. The uses of some applications of the supervisor are shown to reveal their potential. Keywords Smart grid · Advanced distribution management system (DMS) · Virtualization · Supervisor 1 Introduction Historically, electrical distribution networks have not been sufficiently attentive to operational efficiency. Today, the ris- J. B. Leite (B) · J. R. S. Mantovani Electrical Engineering Department, UNESP/FEIS, Ilha Solteira, São Paulo, Brazil e-mail: jonatasboasleite@gmail.com J. R. S. Mantovani e-mail: mant@dee.feis.unesp.br ing demand for electricity, coupled with new technologies, calls for modernization of the distribution system. Future requirements of monitoring and control of the distribution system must become more stringent due to the integrated architecture of the smart grid, which must facilitate data exchange between the DMS and market. For example, recent customer programs of distributed generation and demand response are affecting the operations of the DMS; conse- quently, the massive employment of sensors in the distrib- ution network has become essential to the success of smart grid technologies. Clients of electric utilities consume and generate electricity using schedules and specific rules that require a network with bidirectional energy flow and effec- tive capability of monitoring and control (Glover et al. 2010). The impact of demand response management (DRM) and customer behaviors in particular time periods must be mod- eled or foreseen by the price rules and procedures of reward. These dimensions of modeling can be incorporated into load modeling and forecast algorithms (load flow algorithms, modeling of components of the distribution network, and load and topological forecasts) through direct links between applications of the DMS and DRM. For example, when an application of the DRM tries to release a load in response to an order from the system operator, it needs first to check with the DMS whether the load release causes any violation of con- nectivity, operation, or protection of the distribution network (Niyato et al. 2012). Therefore, the technological advances of the smart grid have a strong influence on the functions of the DMS that must adapt and support all requirements of monitoring and control. The main adaptations are (Fan and Borlase 2009) • Inclusion of optimization for radial and loop-closed net- works, multi-level feeder reconfiguration, multi-objective restoration, and evaluation of future loading; 123 J Control Autom Electr Syst (2015) 26:96–104 97 • Provision of the statistical data of operation and cost to optimize the operation of capacitor banks and voltage reg- ulators in the distribution network; • Consideration of bidirectional power flow in the protection schemes; • Use of load estimation and management of load data in the planning and optimization of network operation; and • Management and forecast of customer behavior in an intel- ligent way. In addition, an advanced DMS must also provide appli- cations and tools of analysis and visualization to incorpo- rate the huge increase of data from field devices. Tools of visualization can provide a detailed and clear vision of the large amount of data; for example, they can show the volt- age magnitude profile with colored contours in the distrib- ution network diagrams, monitor, and alarm voltage devia- tions or show the energy flow in the distribution lines through a contour diagram whose color tones match the line cur- rents. In this way, the system operators and service providers might benefit enormously from these tools in their daily activities. In Part I of this paper, the advanced DMS was designed for the smart grid through a reference model with seven lay- ers: Physical (elements of the power system), Interface (con- nection with physical elements), Communication (allow the data exchange), System (collects data from the communica- tion), Model (abstract representation of the system), Analy- sis (supports the operator’s system decisions), and Intelli- gence (advanced data processing applications). Moreover, the Physical, Interface, and Communication layers were described in depth and mathematically formulated to develop the electrical devices simulator. Here, in Part II of this paper, the computational resources of the System layer are virtual- ized and arranged using the Web-based SCADA architecture; the three logical layers (Model, Analysis, and Intelligence) are then studied and implemented to supervise a real distrib- ution system. 2 Virtualization of the System Layer In the smart grid communication model, the domains of ser- vice provider, grid, and customer are connected through a complex communication network to exchange information among them; thus, the Web-based SCADA architecture pro- vides the necessary resources to implement the smart grid structure. The accessibility of the smart grid services through a typical Web browser allows full interoperability with the traditional Internet Web services, which is the key to success of a global open energy market (Bui et al. 2012). 2.1 Web-Based SCADA Architecture The SCADA system has supported operations of the electri- cal power grid for more than four decades (IEEE PES 2003). Although primitive SCADA systems have been slightly mod- ified and many elements of their architecture are still effec- tive, the traditional SCADA system is not completely effi- cient in operating the smart grid. Fortunately, Internet/Web technologies have been improved for the solution of difficul- ties such as vulnerability to cyber attacks and loss of real-time performance. These technologies, which include a distrib- uted database management system, internet protocol-based multicast communication, and the wide area server structure, increase the reliability of the Web-based SCADA architec- ture (Ebata et al. 2000). In the Web-based SCADA architecture, users are clients who employ browsers to access the Web applications hosted in the servers. Each client is also free to perform any function in the electrical power grid because he does not need any one specific application dedicated to only one task; this increases the flexibility of the SCADA system. Compared to the tradi- tional SCADA architecture, there are other advantages, such as (1) System cost the equipment required to implement a Web- based SCADA architecture is small compared to that required by the traditional architecture. For example, there is no necessity for many local area networks (LANs) or many servers. (2) Data sharing in the advanced DMS, the SCADA sys- tem is often solicited to share data with other comput- ing systems. The Web-based SCADA architecture has the resources that facilitate data sharing via uniform resource locators (URLs), while the traditional architec- ture requires specialized tools to share data, making this operation difficult. (3) Update system when the traditional SCADA system is replaced by the Web-based system, the remote controlled units are reused; each remote unit is connected to the communication network via a gateway that sends and receives controlling signals from the servers. The advantages provided by the Web-based SCADA architecture support the smart grid requirements and clar- ify its usage in the advanced DMS structure. In this way, the main components that constitute the Web-based SCADA architecture, such as Web, databases, and geographical infor- mation servers, are studied and implemented into a virtual environment. 123 98 J Control Autom Electr Syst (2015) 26:96–104 2.2 Development of the SCADA System The physical layer has the equipment of the electrical power grid, while the interface and communication layers are related to the devices of the communication technologies. Thus, the system layer supports the computational machin- ery that runs the various advanced DMS tools employed to process all smart grid information. Figure 1 shows a block diagram of the simulation envi- ronment used to implement the SCADA system and develop advanced DMS tools. The logical blocks represent comput- ers with specific functions. The first represents the electrical devices simulator (EDSIM), which emulates the physical, interface, and communication layers. The second block rep- resents the Web server (WEB SRV), and the third represents the database server (DB SRV). All three computers are virtual machines created over a software layer called VMware ESXi. Each virtual machine has an internet protocol (IP) address and one virtual network adapter connected to a virtual switch that allows communication among the virtual machines. The virtual machines also exchange information with external computers when the host adapter is connected to the inter- net. For example, the virtual machines receive several types of maps from a geographical information system server (GIS SRV) on the internet. The virtualizer, VMware ESXi, was adopted to emulate the SCADA system because it provides many benefits such as free cost of acquisition, capability to create powerful virtual networks, and reduction of the project total cost for the owner (Vmware. Inc 2011). Therefore, the resources of information technology required to operate the advanced DMS can be developed in this low-cost simulation environment. 2.3 Web Server The EDSIM provides real-time metering data from the elec- tric power grid to the system layer; it was formulated and described in Part I of this paper. The Web server is the entrance port to the advanced DMS because the Web-based SCADA architecture is used to manage the distribution system. The Web server has domain name system server (DNS) functionality that converts the computer names to IP addresses. It also has the main function of application server, providing essential technologies such as internet informa- tion services (IIS) that allow distributing and operating Web applications. The Web application used to manage the distribution sys- tem has the characteristics of a webpage whose navigation map is shown in Fig. 2. The homepage is connected to two pages: a download page, which has documents and routines to set up the client browser, and the advanced DMS page, which is the supervisory application and has the EMS functions. The start page displays welcome messages and an initial report Fig. 1 Simulation environment used to develop the advanced DMS of the status of the SCADA system. There are also options to change the initial configurations of the EMS functions, such as select systems, select the map type, visualization options, and the monitoring function. The system editor is an EMS function that allows modi- fying the topology of a selected distribution network or any other selected electrical equipment. The topological modifi- cations are performed using forms that enable the insertion of series- and shunt-connected devices. The dialog window for electrical device status allows verifying the power flow, profiles of voltage and current, and also changing the device status; for example, opening or closing an automatic switch. Events monitoring is another EMS function that permits monitoring the operational conditions of a selected distrib- ution network through tools such as dynamic coloring and an energy-loss locator. The first one displays segments of the distribution network in different color tones according to their voltage profiles, while the second one can indicate electrical devices with energy leakage. Both the system editor and events monitor have multiple layers in their work areas. The front layer shows the schema of the distribution network topology with the symbol of each electrical device; each device location is geo-referenced and shown over a map in the background layer, either a street map or satellite image provided by the GIS server. 123 J Control Autom Electr Syst (2015) 26:96–104 99 Fig. 2 Navigation map of the main webpage hosted at the Web server 2.4 GIS Server Geographical information system technology has been used to improve decisions, manage resources efficiently, and increase the efficiency of work flow, allowing reducing costs for small, medium, and large organizations. When the resources of geographical information are available on the Web, the geospatial data are stored by servers and can be accessed by several users with different permission levels. Moreover, many servers permit the access to the database to perform a consultation or generate thematic maps, as well as any other required function (Booth and Mitchell 2011). There are several servers of geographical information on the Web; the most popular is the Google server. For this project, the ArcGIS server is the adopted provider of geospa- tial data due to its compatibility with the other components of the SCADA system and the availability of a cost-free data- base, which is accessed by the clients as a local database. The ArcGIS server provides many services, such as feature (for web editing), image (for providing control over imagery delivery), and map (for cached and optimized map services), that are employed by the Web application of the advanced DMS in its different layers. For example, the front layer uses the feature service, while the background layer uses the ser- vices of image and map that are automatically updated to facilitate navigation within the distribution network area. 2.5 Database Server The database system is an essential component of the information and communication technologies infrastructure because it stores the data in a systematic manner and enables them to be retrieved, processed, and analyzed either immedi- ately or later; hence, the database server is indispensable for the architecture of the advanced DMS (Shawkat Ali 2013). In the database server, the information is methodically arranged in tables; for example, each electrical device of the distrib- ution system can own a specific table whose columns have values related to its electrical characteristics, control, and status. Figure 3 shows a block diagram of the database server; the node named Main Server represents the virtual machine containing the database of the advanced DMS structure. The database, EMS, is split into two schemas of table groups. The first has the tables of all electrical devices that can operate in the distribution network, such as loads, trans- formers, and lines. The second schema contains the tables that store the metering data from smart meters, PMUs, and IEDs. The database, GIS, and Web servers make up the core of the system layer that represents the last layer of the advanced DMS structure physically equipped with computers; thus, the next three layers are only logical. 123 100 J Control Autom Electr Syst (2015) 26:96–104 Fig. 3 Representation of the database server 3 Model Layer The model layer is the first logic layer and is an abstract representation of the system, communication, interface, and physical layers through standards, such as IEC 61970 and 62325, that establish the rules to represent the topology and metering data of the electrical power grid. 3.1 Common Information Model (CIM) The CIM is standardized by three different IEC standard series: IEC 61970, IEC 61968, and IEC 62325. Each has a particular background and respective focus on energy man- agement system application program interfaces, application integration into electric utilities, and framework for energy market communication (Uslar et al. 2012). Indeed, CIM is an abstract model that represents all of the components in an electric utility, and it defines classes and attributes for these components as well as their relationships using the unified modeling language (UML) notation that can eas- ily be converted into extensible markup language (XML) (Shahidehpour and Wang 2003). Figure 4 shows a comprehensive CIM partitioned into many packages; dependency relationships are indicated by dashed lines with the arrow pointing to the supplier pack- age. IEC 61970-301 defines and describes all packages; for example, • Core contains the basic entities that are shared by all appli- cations; • Topology models the connectivity that is the physical def- inition of how equipment is connected together; • Wires models the information on electrical characteristics of transmission and distribution networks; • Outage models the information on the current and planned network configuration; Fig. 4 CIM top-level packages of the standard IEC 61970-301 • Protection models the information for protection equip- ment such as relays; • Measurements contains the entities that describe exchange of dynamic measurement data between applications; • Load Model provides the models for energy consumers and the system load; • Production provides the models for various types of gen- erators and models the production cost information; • Generation Dynamics provides the models for prime movers needed for simulation and educational purposes; • Domain is a data dictionary of quantities and units that defines data types for attributes that may be used by any class in any other package; • SCADA contains the classes that model the data points located in remote units. Each CIM package contains a number of defined classes that describe the characteristics of a component found in the real electric power system using attributes; for example, Fig. 5 shows the class diagram of a distribution line with some relevant attributes. The Line is found in the wires package that has the class AC Line Segment, with consistent electrical characteristics given by attributes such as susceptance, resis- tance, reactance, and length. The Equipment Container has many types of equipment and is associated with connectiv- ity nodes that are logical topological objects needed to link physical objects together. The attributes of CIM classes provide information from all components of an electric power system to the analysis methodologies such as a state estimator. 4 Analysis Layer The analysis layer has all of the functions and applications to support the operator and automatic decisions using either real-time or historical data from the system layer via the 123 J Control Autom Electr Syst (2015) 26:96–104 101 Fig. 5 Class diagram of the line relevant characteristics model layer. One of main applications of the analysis layer is the state estimation that can obtain good estimates of the state variables by processing the available measurements together with knowledge of the network topology and line model para- meters (Baran and Kelley 1994). The set of network states, �, is composed of the union of the bus voltages, Vi , branch currents, Ji , and current injections, Ii , for m pairs of branch buses along the distribution network, as given by (1). In (2), ψ t i is the set of measured states transmitted to the system layer; their union forms the set of measurements, �. Lastly, the set of unknown states, X, contains all of the network states that are not metered, i.e., is the complement of� in� as expressed by (3). Ω = { ω ∣∣∣∣ω ∈ m⋃ i=1 ([ V̇i ] 3x1 , [ J̇i ] 3x1 , [ İi ] 3x1 )} (1) Ψ = { ψ ∣∣∣∣ψ ∈ Ω � ω ∈ m⋃ i=1 ψ t i } (2) X = {χ |χ ∈ Ω � ω /∈ Ψ } = Ω\Ψ. (3) If a smart metering system monitors the loads of the dis- tribution network, the available measurements must have a large amount data with high accuracy and resolution (Hu et al. 2011). In this way, Leite and Mantovani 2013 propose a rapid and non-iterative state estimation for distribution net- works. It is based on the smart metering system that creates a sequence of calculations for each unknown state using the solution of the Hamiltonian cycle problem found by a search algorithm that minimizes an objective function, F(X). min F(X) = nX∑ i=1 C(χi ). (4) In (4), C (χi ) is the cost of an unknown state; during the search procedure, it becomes zero when the solution equation of χi is found. [ J̇ SE U ] 3x1 = [ İU ] 3x1 + n∑ i=1 [ J̇i ] 3x1 (5) [ V̇ SE U ] 3x1 = [ V̇D ] 3x1 + [ Z̄ D ] 3x3 [ J̇D ] 3x1 (6) [ J̇ SE D ] 3x1 = [ J̇U ] 3x1 − [ İU ] 3x1 − n−1∑ i=1 [ J̇i ] 3x1 (7) [ V̇ SE D ] 3x1 = [ V̇U ] 3x1 − [ Z̄ D ] 3x3 [ J̇D ] 3x1 (8) [ J̇ SE D ] 3x1 = [ n∑ i=1 [ Z̄i ] 3x3 ]−1 ([ V̇U ] 3x1 − [ V̇n ] 3x1 ) . (9) Equations (5) and (6) are solution equations of χi when it is upstream of known states. Similarly, Eqs. (7), (8), and (9) solve χi when it is downstream of known states. These equations are derived from Kirchhoff’s current and voltage laws (KCL and KVL), where U is the index of upstream vari- ables, D is the index of downstream variables, and [ Z̄ ] 3x3 is the matrix of distribution line impedance. Sometimes a dis- tribution transformer replaces a distribution line to connect the medium- to low-voltage networks. [ J̇ SE U (ts) ] 3x1 = K −1 n [ ]3x3 [ J̇D ] 3x1 + [ Ȳm ] 3x3 [ V̇U ] 3x1 (10)[ V̇ SE U ] 3x1 = Kn[ ]3x3 ([ V̇D ] 3x1 + [ Z̄d ] 3x3 [ J̇D ] 3x1 ) (11)[ J̇ SE D ] 3x1 = Kn [ ]T 3x3 ([ Ȳm ] 3x3 [ V̇U ] 3x1 + [ J̇U ] 3x1 ) (12)[ V̇ SE D ] 3x1 = K −1 n [ ]T 3x3 [ V̇U ] 3x1 − [ Z̄d ] 3x3 [ J̇D ] 3x1 . (13) Equations (10) to (13) are adaptations of solution equa- tions for a distribution transformer using the quadrupole model formulation in which Kn is the transformation ratio, [ ]3x3 is the matrix of phase angle deviations, [ Ȳm ] 3x3 is a diagonal matrix of magnetization admittance, and [ Z̄d ] 3x3 is a diagonal matrix of dispersion impedance. In addition, the upstream states are the primary values while the downstream values are secondary values of the distribution transformer. Minimizing F(X) provides a sequence of calculation from the Hamiltonian path that allows the unknown states to be rapidly estimated through a non-iterative method. In this way, the state estimation procedure must solve only the optimiza- tion problem to find a new Hamiltonian cycle when the distri- bution system is alarmed because of a topological change or a fault on AMI, as shown in Fig. 6. In normal operation, state estimation is performed rapidly by the Hamiltonian path that uses an updated set of measurements for each time period. 123 102 J Control Autom Electr Syst (2015) 26:96–104 Fig. 6 Flowchart of the state estimation procedure 5 Intelligence Layer This is the highest layer of the Smart Grid reference model and includes the advanced data processing applications that reduce the need for human intervention as well as provide intelligence to support decision-making. For example, the automated restoration procedure, also known as self-healing, might be deployed on the intelligence layer using remote con- trol and visibility (Cespedes 2012). For self-healing, all pro- tection devices need to have some knowledge of the current network configuration; thus, the localization of the affected points can be readily known during fault events. Self-healing should be integrated with the dynamic col- oring tool that permits to the system operator monitor the energy quality delivered to the end consumers using, for example, the monitoring of the voltage profile. The dynamic coloring tool uses many resources of the advanced DMS such as geo-referenced topological diagrams in multiple lay- ers and the results of state estimation. If the voltage ranges are standardized and associated with a specific color, each branch drawn should have a color in the topological diagram in accord with its voltage value provided by the state estima- tion. Thus, the line model must have a color attribute given by the vector [RG Bi ]1x3 that indicates the quantities of red, green, and blue for the line diagram. [RG Bi ] = ⎧⎨ ⎩ [ 0 1 0 ] , 0.93VB ≤ ∣∣V̇ SE i ∣∣ ≤ 1.05VB[ 0 0 1 ] , 0.90VB ≤ ∣∣V̇ SE i ∣∣ < 0.93VB[ 1 0 0 ] , ∣∣V̇ SE i ∣∣ < 0.90VB � ∣∣V̇ SE i ∣∣ > 1.05VB . (14) In (14), VB is the reference voltage of the network and∣∣V̇ SE i ∣∣ is the bus voltage magnitude at the end of the i th branch. Then the color of drawn branch is either green for an appropriate voltage range, between 93 and 105 % of VB , blue for a poor voltage range, between 90 and 93 % of VB , or red for a critical voltage range, lower than 90 % or higher than 105 % of VB . The dynamic coloring tool updates the coloring of the topological diagram in each time interval, which improves the decision-making process during stress situations. 6 Partial Results and Discussion The computational resources of the System layer were emu- lated using the virtualizer VMware ESXi, where three virtual machines (EDSIM, DBSRV, and WEBSRV) were created and connected via a virtual network (VNLV) as shown in Fig. 7. The networking diagram also demonstrates the con- nection from WEBSRV to the physical adapter of the vir- tual machines server. The physical adapter is connected to an external network with access to the Internet; hence, the WEBSRV was set up as a router to enable communication between external computers and virtual machines in the vir- tual network and vice versa. The WEBSRV virtual machine also has the functions of Web and DNS server; thus, an operator using a workstation connected to an external network can access the supervisory functions of the advanced DMS through its URL address. Figure 8 shows the start page of the supervisor that was accessed using the Internet Explorer browser. The address bar has the URL, and the browser window displays the start page whose welcome message is shown, as well as selected Fig. 7 The Networking diagram of the Smart Grid simulation environ- ment 123 J Control Autom Electr Syst (2015) 26:96–104 103 Fig. 8 Start page of the supervisory of advanced DMS Fig. 9 System Editor application: a feeder selection; b visualization of the selected feeders options of event monitoring, such as dynamic coloring for the phase A. An important application of the supervisor is the Sys- tem Editor that allows the visualization and modification of the distribution network and their electrical devices. On the supervisor main page, one single click of the System Editor tab leads the operator to the application whose work area has a map type from the GIS server and the power substations of the covered region. If any substation is selected, the appli- cation launches a dialog box that shows all feeders of this substation, as shown in Fig. 9a, where feeders #15782 and #24248 are checked. Figure 9b shows the topological dia- gram of the two checked feeders, one black and one blue. The topological diagram is geo-referenced and overlaps the street map provided by the GIS server. It also shows the sym- bols of the electrical devices such as loads, capacitors, and switches. The operator can access electrical device features by dou- ble clicking the symbol of the device of interest. For example, Fig. 9b indicates a switch whose features are shown by the dialog box in Fig. 10. The supervisor and control tab reveals the switch behavior through the monitoring of the electrical parameters such as apparent power, power factor, voltage, and current magnitudes provided by the EDSIM that simu- lates the distribution, interface, and communication layers. This tab also allows the operator to control the switch through the modification of its status to either closed or opened. In Fig. 10, the curve of apparent power indicates the moment when the switch is opened and closed again. The switch stays open for four hours; during this period, the power flow pass- ing through the switch is null. Opening the switch alters not only its power flow but also the electrical parameters along the distribution net- work. Dynamic coloring is the ideal tool to check this event. Because dynamic coloring was the selected option initially, the procedure to access it consists of a single click on the Event Monitoring tab followed by the selection of the sub- station and feeder. Figure 11 shows the results of the dynamic coloring tool during the time period that the indicated switch was open. The section of distribution network upstream of switch remained energized; consequently, its topological dia- gram was green because the estimated voltage of their buses and branches was in the appropriate voltage range. On the other hand, the downstream section was disconnected from energy source; thus, its topological diagram was red because the estimated voltage was zero along this section, and hence it was in the critical voltage range. These results show only some applications of the advanced DMS supervisor; they realize the proposed methodology for the development of a smart grid simulation environment and demonstrate the feasibility to build it. 7 Conclusions The proposed simulation environment provides an alterna- tive way to elaborate new tools of operation, planning, and 123 104 J Control Autom Electr Syst (2015) 26:96–104 Fig. 10 Dialog box of the switch features Fig. 11 Dynamic coloring response to the switch opening control of a smart grid whose framework was designed using a reference model with seven layers; three were mathemati- cally formulated in Part I of this work. The four upper layers are described and implemented in this Part II. The Web-based SCADA architecture was adopted to inte- grate all of the devices and functionalities of the smart grid into a System layer. The computational machinery of the System layer was emulated using the VMware ESXi virtu- alizer that provides benefits such as free-cost distribution, capability to create powerful virtual networks, and reduc- tion of the total cost of the project for the owner. Hence, the proposed methodologies allowed developing a low-cost simulation environment. In addition, the logical layers were integrated by the pow- erful CIM IEC standard; the attributes of its classes provide information about all components of an electric power sys- tem to the analysis methodologies and, subsequently, to the intelligent applications. With this base, the functions of the advanced DMS supervisor that were built and simulated can also be applied to the real world in the future. Acknowledgments This work was fully supported by the Fundação de Amparo à pesquisa do Estado de São Paulo - FAPESP (processos: 2010/07757-1, 2013/23590-9) and CNPq (processo: 305371/2012-6). References Baran, M. E., & Kelley, A. W. (1994). State estimation for real-time monitoring of distribution systems. IEEE Transactions on Power Systems, 9(3), 1601–1609. Booth, B., & Mitchell, A. (2011). Welcome to ArcGIS. Getting started with ArcGIS (1st ed., pp. 5–16). Redlands: ESRI Press. Bui, N., Castellani, A. P., Casari, P., & Zorzi, M. (2012). The internet of energy: a web-enable smart grid system. IEEE Network, 26(4), 39–45. Cespedes, R. (2012). A reference model for the electrical energy system based on smart grids. IEEE T&D LA, 2012, 1–6. Ebata, Y., Hayashi, H., Hasegawa, Y., & Suzuki, K. (2000). Devel- opment of the intranet-based SCADA (supervisory control and data acquisition system) for power system. IEEE Transactions on Power Delivery, 03, 1656–1661. Fan, J., & Borlase, S. (2009). The evolution of distribution. IEEE Power and Energy Magazine, 07(2), 63–68. Glover, J. D., Sarma, M. S., & Overbye, T. J. (2010). Power distrib- ution. Power system analysis and design (5th ed., pp. 759–769). Stamford, CT: Cengage Learning. Hu, Y., Kuh, A., Kavcic, A., & Nakafuji, D. (2011). Real-time state estimation on micro-grids. IEEE IJCNN, 2011, 1378–1385. IEEE PES. (2003). Substation Automation Tutorial (pp. 01–23). Piscat- away, NJ. Leite, J. B., & Mantovani, J. R. S. (2013). State estimation of distribution networks through the real-time measurement of the smart meters. IEEE Grenoble PowerTech, pp. 1–6. Niyato, D., Wang, P., & Hossain, H. (2012). Reliability analysis and redundancy design of smart grid wireless communications sys- tem for demand side management. IEEE Wireless Communication, 19(3), 38–46. Shahidehpour, M., & Wang, Y. (2003). Common information model and middleware for integration. Communication and control in electric power systems: Application of parallel and distributed processing (1st ed., pp. 135–175). New York: Wiley. Shawkat Ali, A. B. M. (2013). Database Systems for Smart Grid. Smart grids: Opportunities, developments, and trends (1st ed., pp. 151– 168). Berlin: Springer. Uslar, M., Specht, M., Rohjans, S., Trefke, J., & González, J. M. V. (2012). The IEC common information model. The common infor- mation model CIM: IEC 61968/61970 and 62325: A practical introduction to the CIM (1st ed., pp. 75–105). Berlin: Springer. Vmware. Inc (2011). Workstation User’s Manual. http://www.vmware. com/support/pubs. Accessed 07 June 2014. 123 http://www.vmware.com/support/pubs http://www.vmware.com/support/pubs Development of a Smart Grid Simulation Environment, Part II: Implementation of the Advanced Distribution Management System Abstract 1 Introduction 2 Virtualization of the System Layer 2.1 Web-Based SCADA Architecture 2.2 Development of the SCADA System 2.3 Web Server 2.4 GIS Server 2.5 Database Server 3 Model Layer 3.1 Common Information Model (CIM) 4 Analysis Layer 5 Intelligence Layer 6 Partial Results and Discussion 7 Conclusions Acknowledgments References