The master's thesis deals with the power management of electric vehicle charging stations in a facility with limited available power capacity. The topic is highly relevant in the context of the ongoing development of electromobility and the increasing demand for efficient utilization of existing electrical infrastructure.

During the preparation of the thesis, the student worked independently and regularly consulted the achieved results. The thesis includes an overview of the basic principles of EV charging management and presents a simulation model of a supply point with limited available power, including a building with a variable load profile and several charging stations.

The main part of the thesis focuses on the design of a power allocation algorithm for charging electric vehicles. The proposed solution is based on a simple priority-based allocation of charging power according to the current state of charge of individual vehicles. Although this approach enables compliance with the power limit of the supply point, I consider its technical contribution to be limited. The algorithm does not take into account several parameters that are essential in real charging management systems, such as vehicle departure times, required energy demand, dynamically changing user priorities, future load predictions, or optimization of the available power utilization.

As a result, the system behavior is largely determined by predefined rules and selected input conditions. The simulation study demonstrates the functionality of the proposed concept; however, it does not provide sufficient evidence to assess its effectiveness in comparison with other approaches commonly used in modern charging infrastructure management systems. The thesis also lacks a deeper analysis of the obtained results and a broader technical evaluation of the proposed solution.

Further reservations concern the consistency and technical depth of several parts of the thesis. A number of assumptions used in the model are introduced without detailed justification, and some conclusions are not sufficiently supported by simulation results. While a functional model demonstrating the basic charging management principle has been developed, the overall level of original technical contribution remains rather limited.

From the perspective of the thesis objectives, the basic requirements of the assignment have been fulfilled. However, considering the simplicity of the proposed algorithm, the limited scope of its validation, the relatively low degree of technical innovation, and the less convincing evaluation of the achieved results, I assess the thesis as meeting only the minimum acceptable standard expected from a master's thesis.

I recommend the master's thesis for defence. Points proposed by supervisor: 55

The student, Ing. Mohammad Rezaul Hoque, MSc, submitted a master’s thesis entitled "Design of Power Management for Electric Vehicle Charging Stations in a Limited Power Supply Facility". The topic is relevant and timely, especially with respect to the development of electromobility, limited reserved power capacities in buildings, and the need for dynamic charging management.

The thesis formally attempts to address the individual points of the assignment. It contains a literature review, a description of a modelled system, a proposed power allocation algorithm, a simulation model, and an evaluation of selected scenarios. A positive aspect is the effort to solve a practical problem: the possible exceeding of the available facility power limit during the simultaneous operation of several EV charging stations. The thesis correctly identifies that unmanaged charging may lead to demand peaks and that dynamic power management may reduce these peaks. The submitted MATLAB file also demonstrates a basic principle of distributing available power among charging points.

However, the technical fulfilment of the assignment is uneven and, in several key aspects, insufficiently convincing. The modelled case is not clearly defined. The thesis does not explain what specific facility is being modelled, why the limit of 86 kW was selected, and whether this value represents a real connection point, contracted reserved capacity, main circuit breaker limitation, or only an assumed simulation parameter. The relation to a three-phase 3×125 A connection is mentioned, but the difference between apparent power and active power is not discussed. Since this limit is the main constraint of the entire simulation, its unclear justification is a serious methodological weakness.

There are also inconsistencies in the description of the building load model. One part of the thesis presents a 24-hour tabular load profile, another describes an analytical curve, while the submitted MATLAB script uses a simplified sinusoidal profile. It is therefore unclear which input data were actually used to obtain the presented results. This is important because the building load directly determines the available power for EV charging.

A major concern is the inconsistency between the thesis text and the submitted computational file. The thesis repeatedly states that a MATLAB/Simulink model was created and describes the implementation as a Simulink architecture. However, only a standalone MATLAB m-file was submitted. No Simulink model, block parameters, subsystem description, or reproducible model documentation are provided. It is therefore not possible to verify whether the declared Simulink model exists or whether the results were generated only by the script.

The thesis also appears to contain unconsolidated versions of the model. For example, the main text and submitted script use an 86 kW limit, while another table states `Pgrid_max = 250 kW`. This discrepancy is not explained, although the maximum available power is the key constraint of the whole work. The chapter "Revision 2 – Response to Supervisor Feedback" is also inappropriate for a final thesis. Such comments should be incorporated into the relevant chapters, not left as a separate change-log-like section. Moreover, this chapter mentions several advanced elements, such as OCPP 2.0.1, Modbus TCP, realistic load modelling, battery models, CC-CV charging, dynamic priority indices, and economic analysis, but it is unclear whether these were actually implemented. The submitted code corresponds to a much simpler model.

After checking the submitted "EV_Simulation_Main.m" file, I conclude that it is only a simple discrete-time MATLAB simulation. It uses four EVs, fixed battery capacities, fixed arrival and departure times, fixed priorities, a fixed 86 kW import limit, an 8-hour simulation interval, and a sinusoidal building load. The priority-based algorithm only sorts active vehicles by fixed priority and allocates available power sequentially. It does not consider remaining time to departure, required energy, dynamic priorities, minimum charging currents, phase constraints, communication delays, or an optimisation objective. The “Equal Sharing” strategy also does not redistribute unused power when some vehicles reach their maximum charging power.

The results are therefore largely predetermined by the implemented rules. The controlled strategies cannot exceed the grid limit because they are directly constructed from the available power limit. The uncontrolled strategy ignores the limit and therefore exceeds it under sufficient simultaneous demand. The conclusion that dynamic load management prevents limit violations is thus more a consequence of the implemented rule than a result of a validated and independently verified model. The code can be considered a basic demonstration of the principle, but not a sufficiently documented, validated, and reproducible computational model for a master’s thesis.

The formal and textual quality of the thesis is also weak. Many chapters consist mainly of bullet points, short statements, and general lists without deeper technical discussion. The literature review often appears as a compilation of general information rather than a critical evaluation leading to the proposed solution. Figures and tables are not consistently formatted, numbered, explained, or connected to the text. Some tables overflow beyond the page margins, and several figures merely repeat information from bullet points without adding technical value. Some diagrams are difficult to read due to overlapping arrows and text.

The thesis also shows signs of insufficiently checked machine-generated or machine-translated content. The exact origin of the text cannot be proven from the document alone, but several formulations are technically or linguistically inappropriate. Examples include unrelated or awkward phrases in technical sections, which indicate insufficient authorial control and editing. The possible use of generative AI or machine translation is not problematic in itself if allowed, transparently declared, and carefully checked. However, if such tools were used without proper declaration or verification, this would be a serious issue regarding authorial responsibility and trustworthiness.

The author’s own contribution can be identified mainly in the creation of a basic MATLAB simulation script and a simple charging power allocation rule. However, this contribution is weakened by insufficient documentation, unclear originality, missing validation, and inconsistencies between the text and the submitted implementation.

Overall, the thesis addresses a relevant topic, but in its submitted form I do not consider it to meet the expected requirements for a master’s thesis. The most serious deficiencies are the unclear definition of input parameters, inconsistent load modelling, the discrepancy between the declared Simulink model and the submitted MATLAB script, different values of key parameters, the inappropriate inclusion of the revision chapter, weak formal quality, insufficient discussion, missing validation, and an insufficiently convincing demonstration of the author’s own contribution.

For these reasons, I recommend the thesis for defence only with very serious reservations. In its submitted form, I propose the following assessment:

Proposed assessment: F / failed – 45 points.

Only if the student convincingly demonstrates during the defence which model is final, how the results were obtained, provides the actual Simulink model or clearly explains its absence, proves reproducibility of the results, and transparently explains any use of generative AI or machine translation, could a conditional assessment at the lowest boundary of grade E be considered. Topics for thesis defence:

1. Explain the origin of the 86 kW power limit. Is it a contracted reserved capacity, a value derived from a 3×125 A main circuit breaker, a real facility parameter, or only a selected simulation parameter? If the value is derived from a three-phase connection, explain how the difference between apparent power, active power, and power factor was considered.
2. The thesis presents several different descriptions of the building load profile: a 24-hour tabular profile, an analytical curve, and a sinusoidal curve used in the submitted MATLAB script. Which building load model was actually used to obtain the results presented in the thesis?
3. The thesis states that a MATLAB/Simulink model was created. However, only a standalone MATLAB m-file is provided as an appendix. Present and explain the actual Simulink model, or clarify why it is not included in the appendices and whether the results in the thesis were obtained from Simulink or only from the submitted script.
4. Explain why different values of maximum available power appear in the thesis, especially 86 kW and 250 kW. Which value was actually used for the results presented in the thesis and why?
5. Did the author declare the use of generative artificial intelligence tools or machine translation in the preparation of the text, figures, or other parts of the thesis? If so, to what extent were these tools used, and how was the technical correctness of the outputs verified?
6. What exactly is the author’s own contribution compared to standard dynamic power management of EV charging stations?
7. How was the simulation model validated? Is there any comparison with real data, a reference model, or another computational tool?
8. When an EV reaches the target SoC, the MATLAB script clips the SoC to the target value, but does not reduce the allocated power in the final time step according to the remaining energy demand. How does this affect the calculated delivered energy and charging time?

Points proposed by reviewer: 45