Ioannis Konstantelos

The ongoing electrification of the transport sector is expected to cause an increase in electricity demand and, therefore, trigger significant network investments to accommodate it. This paper focuses on investment decision-making for electricity distribution grids and specifically on the strategic and incremental investment network planning approaches. In particular, the former involves network planning with the consideration of a long-term multi-stage study horizon, as opposed to a shorter–term view of the future that applies to the latter case. An investment analysis that is carried out underlines the economic savings generated from adopting a strategic investment perspective over an incremental one. These economic savings are achieved from the fact that the associated fixed investment costs are incurred only once in the horizon under strategic planning. On the other hand, incremental planning involves a series of network reinforcement decisions, thereby incurring the fixed cost multiple times. In addition, sensitivity analyses that are carried out capture the effect of key parameters, such as investment cost, discount rate and investment delay, on the generated economic savings.

The decarbonisation of the electricity grid is expected to create new electricity flows. As a result, it may require that network planners make a significant amount of investments in the electricity grids over the coming decades so as to allow the accommodation of these new flows in a way that both the thermal and voltage network constraints are respected. These investments may include a portfolio of infrastructure assets consisting of traditional technologies and smart grid technologies. One associated key challenge is the presence of uncertainty around the location, the timing, and the amount of new demand or generation connections. This uncertainty unavoidably introduces risk into the investment decision-making process as it may lead to inefficient investments and inevitably give rise to excessive investment costs. Smart grid technologies have properties that enable them to be regarded as investment options, which can allow network planners to hedge against the aforementioned uncertainty. This paper focuses on key smart technologies by providing a critical literature review and presenting the latest mathematical modelling that describes their operation.

The building sector has traditionally accounted for about 40% of global energy-related carbon dioxide (CO₂) emissions, as compared to other end-use sectors. Due to this fact, as part of the global effort towards decarbonization, significant resources have been placed on the development of technologies, such as active buildings, in an attempt to achieve reductions in the respective CO₂ emissions. Given the uncertainty around the future level of the corresponding CO₂ emissions, this work presents an approach based on machine learning to generate forecasts until the year 2050. Several algorithms, such as linear regression, ARIMA, and shallow and deep neural networks, can be used with this approach. In this context, forecasts are produced for different regions across the world, including Brazil, India, China, South Africa, the United States, Great Britain, the world average, and the European Union. Finally, an extensive sensitivity analysis on hyperparameter values as well as the application of a wide variety of metrics are used for evaluating the algorithmic performance.

Given its physical characteristics and the range of services that it can provide, energy storage raises unique modeling challenges. This paper summarizes capabilities that operational, planning, and resource-adequacy models that include energy storage should have and surveys gaps in extant models. Existing models that represent energy storage differ in fidelity of representing the balance of the power system and energy-storage applications. Modeling results are sensitive to these differences. The importance of capturing chronology can raise challenges in energy-storage modeling. Some models ‘decouple’ individual operating periods from one another, allowing for natural decomposition and rendering the models relatively computationally tractable. Energy storage complicates such a modeling approach. Improving the representation of the balance of the system can have major effects in capturing energy-storage costs and benefits.

This paper focuses on developing an appropriate combinatorial optimization technique for solving the optimal sizing problem of hybrid AC/DC microgrids. A novel two-stage iterative approach is proposed. In the first stage, a metaheuristic technique based on a tailor-made genetic algorithm is used to tackle the optimal sizing problem, while, in the second, a non-linear solver is deployed to solve the operational problem subject to the obtained design/investment decisions. The proposed approach, being able to capture technical characteristics such as voltage and frequency through a detailed power flow algorithm, provides accurate solutions and therefore can address operational challenges of microgrids. Its capability to additionally capture contingencies ensures that the proposed sizing solutions are suitable both during normal operation and transient states. Finally, the genetic algorithm provides convergence of the model with relative computational simplicity. The proposed model is applied to a generalizable microgrid comprising of AC and DC generators and loads, as well as various types of storage technologies in order to demonstrate the benefits. The load and natural resources data correspond to real data.

The ongoing decarbonisation of modern electricity systems has led to a substantial increase of operational uncertainty, particularly due to the large-scale integration of renewable energy generation. However, the expanding space of possible operating points renders necessary the development of novel security assessment approaches. In this paper we focus on the use of security rules where classifiers are trained offline to characterize previously unseen points as safe or unsafe. This paper proposes a novel deep learning-based feature extraction framework for building security rules. We show how deep autoencoders can be used to transform the space of conventional state variables (e.g., power flows) to a small number of dimensions where we can optimally distinguish between safe and unsafe operation. The proposed framework is data-driven and can be useful in multiple applications within the context of security assessment. To achieve high accuracy, a novel objective-based loss function is proposed to address the issue of imbalanced safe/unsafe classes that characterize electricity system operation. Furthermore, an R-vine copula-based model is proposed to sample historical data and generate large populations of anticipated system states for training. The superior performance of the proposed framework is demonstrated through a series of case studies and comparisons using the load and wind generation data from the French transmission system, which have been mapped to the IEEE 118-bus system.

Supervised machine learning has been successfully used in the past to infer a system's security boundary by training classifiers (also referred to as security rules) on a large number of simulated operating conditions. Although significant research has been carried out on using classifiers for the detection of critical operating points, using classifiers for the subsequent identification of suitable preventive/corrective control actions remains underdeveloped. This paper focuses on addressing the challenges that arise when utilizing security rules for control purposes. The inherent trade-off between operating cost and security risk is explored in detail. To optimally navigate this trade-off, a novel approach is proposed that uses an ensemble learning method (AdaBoost) to infer a probabilistic description of a system's security boundary and Platt Calibration to correct the introduced bias. Subsequently, a general-purpose framework for building probabilistic and disjunctive security rules of a system's secure operating domain is developed that can be embedded within classic operation formulations. Through case studies on the IEEE 39-bus system, it is showcased how security rules can be efficiently utilized to optimally operate the system under multiple uncertainties while respecting a user-defined cost-risk balance. This is a fundamental step towards embedding data-driven models within classic optimisation approaches.

This paper is concerned with the design of an autonomous hybrid alternating current/direct current (AC/DC) microgrid for a community system, located on an island without the possibility of grid connection. It is comprised of photovoltaic (PV) arrays and a diesel generator, AC loads, and battery energy storage devices for ensuring uninterruptible power supply during prolonged periods of low sunshine. A multi-objective, non-derivative optimisation is considered in this residential application; the primary objective is the system cost minimisation, while it is also required that no load shedding is allowed. Additionally, the CO₂ emissions are calculated to demonstrate the environmental benefit the proposed system offers. The commercial software, HOMER Pro, is utilised to identify the least-cost design among hundreds of options and simultaneously satisfy the secondary objective. A sensitivity analysis is also performed to evaluate design robustness against the uncertainty pertaining to fuel prices and PV generation. Finally, an assessment of the capabilities of the utilised optimisation platform is conducted, and a theoretical discussion sheds some light on the proposal for an enhanced design tool addressing the identified issues.