*5.2. Modeling the System-To-Be*

We first need to identify the target problem and its relation to the execution environment to model the system-to-be. From the analysis of the system-as-is, we understand that a possible problem is related to the excessive energy consumption due to the advantages that electricity offers rather than to the technical energy losses related to the end-users' alienation from the energy provision. Energy stealing was also a problem, motivated by changes in the user attitude.

The system's environment does not a general provider, turning it into an isolated microgrid very dependent on the energy it could generate and responsible for its own backup to avoid interruptions.

We propose changing the domain from an isolated microgrid to a distributed microgrid (on-grid) with a higher degree of automation to solve these problems. Domain changes, such as automated measurement and new communication technologies, appear as new objectives. In addition to technical changes, the new system must be safe and reliable, and must meet all interested parties' requirements (such as the utility, financial, and environmental agencies). In the new scenario, the user becomes a prosumer and should provide information feedback to optimize the microgrid system.

The new system can no longer be considered a "product". Therefore, it will be necessary to prioritize the requirements of the system that provides this service.

The primary (system) goal of the system-to-be is *to improve the low voltage legacy microgrid design in the Amazon region*. Still, now the objective is to enhance the automation level. Emerging requirements anticipated in the system-as-is modeling appear in expectations, obstacles, or restrictions, and lead to new requirements.

Thus, to achieve the System To-Be's objectives, we propose a SoS+ architecture that supports a new arrangement of services (subsystems) where agents are dynamic elements. The final user is one of these dynamical elements and has a complex structure of resources, as shown in Figure 13. One of these constituent systems is the smart meter (SM).

The system-as-is goal showed that the power supply would be blocked whenever the user consumption exceeds a pre-established limit of 100 kWH. There is no consideration of historical consumer data.

Figure 14 shows the objective diagram for the system-to-be, in which the main objective is to "Improve the low Voltage legacy Microgrid in the Amazon region". The sub-objectives "Improving Strategic Planning", "Optimize energy production", and "Improve operational management" should contribute to the achievement of this primary goal. These sub-objectives can be successively refined.

The analysis of the objective diagrams points to the need for a service system that supports a higher level of control and interaction with the prosumer. Parameters such as the energy production and consumption in each unit should be monitored, as well as its balance between provision and consumption. All data should be directed to an advanced metering infrastructure (AMI) which acts as a service provider instead of a passive supervisor. The detection of electricity stealing is a goal. Naturally, the SM should process information and commands to activate or deactivate the dealership's energy provision.

The goal of "Reducing energy consumption" would fail if the user does not take the actions required. These actions should range from energy-saving acts to the production of energy as a prosumer. The adequate coupling with the end-user is ensured by monitoring the transaction communication.

**Figure 13.** User domain.

#### 5.2.1. The Object Model

An object model should describe all the agents that contribute to the functioning of the system service, be it humans or machines. Object modeling is also practiced in other representations, including UML, and does not constitute a novelty. It is also essential for the goal-oriented approach to support the interaction between the energy system service and its prosumers. We will further illustrate the modeling of the role of a machine, namely the smart meter.

The prosumer characteristic appears in the goal-oriented model, including the same individual prosumer as an external user (and therefore a consumer) and as part of the system (an energy service provider).

Agent modeling will also be used to support responsibility diagrams.
