PhD defence by Babak Arbab-Zavar


17.08.2022 kl. 13.00 - 16.00


Babak Arbab-Zavar, AAU Energy, will defend the thesis "Communication Architecture Designs of Smart Inverters for MGs"


Communication Architecture Designs of Smart Inverters for MGs


Babak Arbab-Zavar


Professor Josep Guerrero


Professor Juan C. Vasquez


Associate Professor Sanjay Chaudhary


Associate Professor Szymon Beczkowski, Aalborg University, Denmark (Chairman)
Associate Professor Samir B. Belhaouari, College of Science and Engineering, Hamad Bin Khalifa University, Doha, Qatar
Professor Gabriel Garcerá Sanfeliú, UPV Universitat Politècnica de València, Spain


The microgrid (MG) concept was established on the idea of localizing energy generation and consumption while utilizing as many renewable resources as possible. In this respect, the dependence on centralized generation running on conventional fossil fuels is minimized and extensive losses of large distribution systems can be avoided. In MG scenarios, since localizing
is the key, machine-to-machine (M2M) communications are tended to be minimized.
However, M2M communication frameworks are still required, and this PhD thesis is about investigating this problem and offering solutions that can fulfil the requirements and overcome existing issues from a new perspective.
In this thesis that is derived from four journal papers (three published and one submitted), several issues related to the role of communication in MG management are addressed and discussed in detail.

At the first step, specifications of smart inverters were introduced and conceptualized. These power-electronics-based devices act as power interfaces between distributed generators (DGs), loads, and energy storage systems (ESSs) with the MG bus. Properly designed and implemented M2M communication architectures are one of the main requirements of smart inverters
which were explained and justified in this step.
At the next part, a communication architecture was proposed that may successfully be employed into grid-connected (GC) MG systems to properly control the storage unit and minimise energy trades with the stiff grid at the point of common coupling (PCC). In this part, the communication network was fully developed from scratch by using micropython-enabled microcontrollers
that are low power devices and can support the internet of things (IoT) ideology. Moreover, the message queuing telemetry transport (MQTT) application protocol was selected for messaging according to its lightweightness and publish/subscribe methodology, therefore the whole design was aiming to reduce energy expenditure and complexity. Alongside the communication framework design, the MG model formed on grid-feeding (GFD) inverters was also designed and implemented on an experimental testbed, and results were provided to validate this integration.

The physical layer protocol of the aforesaid communication architecture was short-range (WiFi). This issue might cause limitations if the specific MG of interest contains geographically widespread units which motivated the
next part of this thesis. Long-range protocols were studied and specifically low-power, wide area network (LPWAN) technologies were considered as a solution to the coverage problem. Among the various protocols and systems LoRaWAN was selected owing to its range, implementation simplicity, and possibility to construct private networks. In this accord, a LoRa enabled
smart inverter model was prepared and experimentally evaluated. Such a smart inverter can connect widespread DGs, or ESS with each other or an MG central controller (MGCC) in an effective and dependable fashion.
In the next part of this thesis, the effects of transmission link-loss on MG optimum performance were explained and investigated. Knowing that depending on the MG architecture and the kind of power-converters used, the communication function on management and control paradigms differs, the identical MG structure of the earlier parts of the study was used here again
for the sake of consistency. In this section, short-term load/generation forecasting methodologies based on deep learning (DL) techniques were performed on synthetically generated databases. These forecastings were used to assist the energy management system (EMS) to anticipate future time-step power setpoints for the storage unit. Consequently, the battery is loaded with supplementary information that can help to mitigate the detrimental repercussions of communication link-losses.
As final remarks and to provide a brief summary of the work, it may be said that in this PhD project the role of communication on AC-MG management was investigated from both communication science and control engineering
viewpoints. This represents a non-trivial and challenging approach that require deep understandings of both areas of technology. Device-level models for control and messaging systems were developed, the integration methodology was presented, and different operational scenarios were evaluated and experimentally validated. The reason that both of the aforesaid
points of view are included is the factor that can distinguish this research from similar studies.


THE DEFENCE will be IN ENGLISH - all are welcome.




See the list