*3.1. Reference Architecture*

Smart grid software architecture is being explored by numerous researchers, service providers, and electricity producers. These works are aimed towards all stakeholders to accelerate smart grid component development and comply with uniform protocols. A reference model is a set of views and descriptions that provides the basis for discussing the characteristics, uses, behavior, interfaces, requirements, and standards of the smart grid [5]. Reference architecture provides a technology strategy that will holistically leverage standards-based solutions to realize these benefits consistently across multiple stakeholders [17–19]. Table 3 summarizes the various standard architectures found in the literature.

FREEDM architecture [24,25,28] facilitate seamless integration of distributed renewable power generation and storage resources and describes three states of operation, namely user level, solid state transformer level, and system level. FREEDM is an appropriate architecture if all of the hardware nodes and communication protocols are homogeneous, whereas CosGRID [18] architecture aims at smart grids based on different hardware platforms and protocols. CoSGRID architecture provides a generic platform suitable to develop improved electric power managemen<sup>t</sup> services and applications for any environment and provide core smart grid control mechanisms. ScorePlus architecture [21] combines the merits from both hardware and software platforms to examine the performance of Smart Grid applications under realistic communication and computation constraints.


### **Table 3.** Reference Architecture.

GridOPTICS architecture [23] is layered architecture that addresses the data encapsulation and dependencies between layers. This promotes a clear separation of concerns and makes it possible to independently develop capabilities within each layer while ensuring data driven mechanisms. Smart grid architecture based on IEC 61499 and IEC 61850 has been experimented in previous studies [29,30]. IEC 61850 decomposes power substations, including functions for monitoring, control, protection and primary devices, down to objects, thus obtaining an object-oriented representation of the power system.

Duke Energy platform [18] implements an electric grid with "distributed intelligence" that significantly increases the operational efficiencies of the electric power system. It uses OSI model and Internet Protocol Suite as a reference to abstract and differentiate the connectivity and computing requirements. This architecture provides routing, bridging, and gateway capabilities to the IP based networks. NASPInet [27] architecture reflects the hierarchy among power grid operators, monitors, and regulators in order to provide a realistic deployment path.

Building Energy Management Open Source Software (BEMOSS) [22] is an open source operating system that is expected to improve sensing and control of equipment to reduce energy consumption. This focusses on user managemen<sup>t</sup> in addition to the operating system and connectivity [31]. BEMOSS is an open source architecture, which enables developers with different skillsets to work on different layers, thus enabling rapid deployment.

### *3.2. Layered Architecture in Smart Grid Software*

The software architecture is a very high-level design of large software systems and the overall structure of a system with a set of defined rules. NIST recommended a conceptual model [5] for smart grid architecture, which is considered as reference architecture, including five domains pertaining to three areas: Smart Grid Service and Applications, Communication, and Physical Equipment.


The NIST architecture model is represented in Figure 2, interconnecting the domains and service areas. NIST model enables the smart grid architecture to be flexible, uniform, and technology-neutral. This explains the simplified domain model in ICT perspective and the mapped domain model on the NIST smart grid interoperability framework. It is meant to foster understanding of smart grid operational intricacies and describes the data flow techniques amongs<sup>t</sup> the participants.

Service provider domain (markets, operators, service providers).

**Figure 2.** NIST reference model for smart grid architecture [32].

The NIST conceptual model revolves around the communication network and the information exchange within grid domains. The interoperability of the domain is represented with numbers (Port 1 to 5) in Figure 2. The insight of different architecture layers (in literature survey) is described under each communication port.

**Communication Port 1:** Describes information exchange between grid domain and communication network. This includes exchange of control signals between devices in grid domain and the service provider domain.

The data and control signal interactions are enabled with the physical layer [33], which consists of thousands of data points. IEDs have real-time capabilities for capturing and processing massive amounts of information of large-scale smart grids. This type of layer brings a component that translates the different communication protocols and data models to the common communication technology and information model. The GTDU layer [3] implements key grid functionalities, such as distributed V-AR (Volt Ampere Reactive) compensation schemes, estimating generation schedules, and regulating power consumption.

In Gridstat architecture [34], data plane components [35] are designed to transmit the data from each source to many potential destinations, as directed by management-plane components. MultiFLEX architecture [36] implements the interoperability across heterogeneous devices using middleware technologies.

System Control Layer [36] represents the coordinated control required to meet the functional and performance-system-level objectives, such as coordinated volt-var regulation, loss minimization, economic and secure operation, system restoration, etc. The system control layer continuously monitors the system devices and keeps track of the system state by using applications such as state estimators. The sensor and actuator layer [37] ensure the local data processing amongs<sup>t</sup> grid components.

Pérez et al. applied autonomic computing techniques towards a reference architecture for micro grids [14]. The autonomous behavior helps real time capabilities to acquire, distribute, process, analyze, connect, and disconnect distributed resources. This model implements the grid controller and the real time scenarios of all participating elements

**Communication Port 2:** Enables metering information exchange between the smart metering domain and communication network, which is a vital element in smart grid architecture. It also warrants the exchange of metering information and interactions through operators and service providers.

Smart meter communication is achieved by the storage layer [34] that collects the data from meters and devices in gird. The component layer [33] analyses data and behavior and provides intelligence to the power network. The data acquisition, processing, and analysis operations are performed on the information received from IED. The connectivity layer [22] takes care of the communication between the operating system and the framework layer and all physical hardware devices. This encompasses different communication technologies, data exchange protocols, and device functionalities.

The Meter Data Interface Layer [38] integrates smart meter architecture with Distribution Management Systems (DMS). A novel AMI communication architecture [39] is described with a Local Metering Concentrator layer. Such an architecture ensures command data processing and data acquisition. Meter Data Interface layers [38–40] perform: (i) pushing of AMI meter data to DMS; (ii) polling of DMS meter data from the AMI; and (iii) pushing of DMS control commands to AMI [41,42].

**Communication Port 3**: Enables the interactions between operators and service providers in the service provider domain and devices in the customer domain. The interface between the smart grid and the customer domain is of special importance as the most visible and user-focused interactive element. Electricity usage is measured, recorded, and communicated through this port.

The service layer [7] consists of those services that are provided to the stakeholders of each of the systems that constitute the large-scale smart grid. The customer services, such as stake on pricing and control of equipment, are implemented under customer-level control mechanisms [3]. The emery utilization is optimized by the information available, such as peak load, climate conditions, and power consumption. The local control layer [3] is used to implement the consumer level control functionalities and demand optimization.

The market layer [36] utilizes all the system control information and uses advanced economic and financial applications, such as reserve co-optimization, risk management, load and price forecasting, and architecture to reduce distribution losses [43] and is designed to coordinate between market dynamics and regional load despatch center.

**Communication Port 4**: Enables information exchange between the service provider domain and communication network domain. It enables communication between services and applications in the service provider domain to actors in other domains.

The service provider communication is enabled in M2M layer [44]. It increases the scalability because it removes the interdependencies between producer and consumer related to the information, allowing the development of services completely independent from the systems and deployed devices. Orchestration layer [33] represents those services that are implemented by the composition of simple services that are required by service providers.

The market (business) layer deals with the process of control decisions for the available resources, incorporating economic objectives [36,45]. This layer enables the service providers to participate in grid level transactions. The service provider domain operations, such as retail energy options, billing, etc., are described with utility provider and third-party provider services [46]. The protocol vulnerability risks for substation and transmission asset mitigation assessments are also addressed.

**Communication Port 5**: Enables the communication between smart metering and the customer domain. The coordination layer [3] defines the information exchanged between the GTDU layer and smart meter to control and optimize functionality implementation. The security layer [36,47] enables secure and trusted communication to control whether a device or service can be trusted or not. The mechanism triggers mutual authentication by providing the means to create a public key infrastructure.

In the MDI layer [38], translation means converting the AMI data to the DMS data when meter data from the AMI is delivered to the DMS, and vice versa. Service-oriented middleware [36] architecture is designed to support different kinds of smart meters and distributed power consumption information. Additionally, this architecture provides transparent interpretation of application and end users by controlling information exchange flow.

### Comparative Study on Different Architectural Layers

Table 4 summarizes the comparison of layered architecture available in various literature. Analysis of various architectural features and key attributes are presented.


### **Table 4.** Comparison of layered architectures.

