Electricity demand forecasting in industrial and residential facilities using ensemble machine learning

Changes to the electricity supply industry, in particular the progress towards deregulation, have prompted an increased interest in short-term electricity demand forecasting. Trend removal has been identified as a key element in forecasting, and perhaps the area that can most benefit from the application of advanced time series analysis. A new approach to trend removal is presented based on fixed interval, or optimal, smoothing techiques. Forecasts derived in this way are compared with more conventional approaches. Preliminary results indicate the new method to be promising and deserving of further development.

Seasonality methods have been developed to model the intraday, intraweek and intrayear seasonal cycles of the electricity load data in one-day ahead electricity demand forecasting. In this paper, we investigate the short-term modeling and forecasting of electricity demand where an intramonth cycle has also been discovered. Thus based on the intramonth cycle, a new mathematical modeling scheme is developed for HWT exponential smoothing model to accommodate the intramonth seasonal cycle and by using six years of Singapore data. We show that fourfold seasonal method outperforms the triple seasonal method in Singapore.

In this paper, we present a novel nonparametric algorithm for short term electricity demand forecasting. The algorithm is based on local linear regression using sliding window with variable length. The method for selecting optimal window length for each local fit offers close insight into trade-off between bias and standard deviation of local regressions. Optimal window length is selected for each value in the load time-series: large window for linear change of load to reduce variability and small window when load departs from linear function to control bias. In the presented algorithm local linear regression is used to estimate trend component of the load time series and to forecast trend component by extrapolating with the fitted local linear function. Some features of the algorithm are demonstrated in the paper using examples from the historic load data recorded in the Namibian Power Utility.

As the development of the Internet of Things, many terminal devices are continually intelligent. However, with the increase of distributed terminal equipment and the diversification of business, cloud computing technology cannot meet the needs of smart home. In consideration of the advantages of edge computing and the demand of smart home, this paper proposes an smart home electricity demand forecasting system based on Edge Computing with a short-term (24-steps) electricity demand forecasting. Our system utilizes intelligent home gateway to store heterogeneous data into a central repository where it will be processed and analyzed, and these analyzed data would then be used for forecasting local bundled resident electricity demand mainly at intelligent home gateway by acting as a local computing unit to provide edge computing service for residents. Besides, our system can also report the historical and real-time environmental data (indoor and outdoor) and electric data from intelligent home gateway. The experimental results demonstrate our system can provide better service quality and scalability with limited computing resources compared with simply using cloud computing system only.

Short-term electricity demand or load forecasting is the way of estimating future demand for short time horizon, an hour to one week ahead. Short horizon forecasting helps to maintain secure power system, avoids blackout risk and provides adequate electricity supply. Therefore, accurate forecasting is prime concern for day to day planing and ensuring the stability of electricity system. Many literature show their mean absolute percentage error (MAPE) for short term load forecasting as less than 2%, which is very competitive performance. The methodology behind good performance is that the model should address the driving factors that effect electricity load and employs appropriate estimation techniques. In our work, multiple linear regression model are developed and Bayesian estimation technique is used. Assuming historical load, temperature, daily, and weekly seasonal patterns as the main effecting factors for electricity load consumption, two models Model A, and Model B are constructed. Model A consist historical load and deterministic terms. However, model B consist all variables of Model A plus temperature variables. These models are analyzed based on their forecasting performance for three years out of sample prediction. Our results showed that when temperature variable is included in model, it can improve the overall performance at least by 20%. Performance improvement was higher during morning and evening hours than day hours. Interestingly, evening hours (17 to 22) are statistically not significant.

The issue of nature management and environmental protection is one of the important areas of sustainable development of society and requires an integrated approach at different levels of management. One of the largest polluters of the natural environment is the construction industry. According to scientific research, construction has led to 50% of climate change. And this process is gaining momentum. The construction industry faces quite responsible tasks of restoring residential, industrial and production facilities. This needs to be addressed both urgently and, especially, after the cessation of hostilities. Russia's military aggression against Ukraine has caused enormous physical damage, but even greater, which cannot even be estimated, moral damage – the lives and dreams of Ukrainians have been destroyed. The entire process of construction and installation works has a significant impact on the natural environment – this includes cutting the vegetation layer during excavation work; changing the relief; harmful emissions into the air, water, soil; waste generation during construction. The upper fertile layer of the earth is also destroyed in the territories that are used temporarily during construction. When performing earthworks, most of the developed soil goes into dumps, which negatively change the natural landscapes, contribute to the erosion of the soil cover; work construction machines is accompanied by toxic exhaust gases, noise, vibration, dust formation. In addition, the construction materials themselves can negatively affect the environment – this is radioactivity, toxicity. Practitioners emphasize the need to define the development of construction materials as one of the priority sectors of the domestic economy. Reforming the regulatory framework in construction, including pricing issues, also includes addressing issues related to environmental management. To optimize environmental management in the construction industry, it is necessary to take into account the capabilities of industries related to construction and individual enterprises. Even at the design stage of construction projects, it is necessary to make environmental choices - choosing building materials without harmful emissions both during their manufacture and during operation; choosing modern equipment with lower noise and vibration levels during construction work; minimizing construction waste and providing for its reuse. After the end of hostilities, Ukraine has a chance to rebuild its residential, public and industrial facilities, taking into account all the principles of “green construction”.

Optimization of wastewater treatment systems for growing industrial parks

In this paper, a systematic literature review is presented, through a survey of the main digital databases, regarding modelling methods for Short-Term Load Forecasting (STLF) for hourly electricity demand for residential electricity and to realize the performance evolution and impact of Artificial Intelligence (AI) in STLF. With these specific objectives, a conceptual framework on the subject was developed, along with a systematic review of the literature based on scientific publications with high impact and a bibliometric study directed towards the scientific production of AI and STLF. The review of research articles over a 10-year period, which took place between 2012 and 2022, used the Preferred Reporting Items for Systematic and Meta-Analyses (PRISMA) method. This research resulted in more than 300 articles, available in four databases: Web of Science, IEEE Xplore, Scopus, and Science Direct. The research was organized around three central themes, which were defined through the following keywords: STLF, Electricity, and Residential, along with their corresponding synonyms. In total, 334 research articles were analyzed, and the year of publication, journal, author, geography by continent and country, and the area of application were identified. Of the 335 documents found in the initial research and after applying the inclusion/exclusion criteria, which allowed delimiting the subject addressed in the topics of interest for analysis, 38 (thirty-eight) documents were in English (26 journal articles and 12 conference papers). The results point to a diversity of modelling techniques and associated algorithms. The corresponding performance was measured with different metrics and, therefore, cannot be compared directly. Hence, it is desirable to have a unified dataset, together with a set of benchmarks with well-defined metrics for a clear comparison of all the modelling techniques and the corresponding algorithms.

Anchorage areas are essential for safe and efficient maritime operations. However, conventional forecasting models often underperform in dynamic port conditions, as they rely heavily on historical averages and static assumptions. To address these limitations, this study proposes a forecasting framework for anchorage occupancy. This framework uses stacked ensemble learning, integrating both statistical and machine learning models to enhance predictive accuracy and operational reliability. The proposed approach was applied to occupancy data from the E1 anchorage at Ulsan Port, with performance evaluated across various forecasting models and ensemble strategies. In addition, a hexagon-based occupancy estimation method was implemented to assess spatial efficiency and safety in comparison to the traditional anchor circle method. The results demonstrate that the stacking ensemble model effectively captures complex, nonlinear patterns in vessel traffic and delivers improved forecasting performance. These findings highlight the practical potential of stacking ensemble techniques and spatial modeling innovations in enabling proactive anchorage management, reducing congestion, and enhancing maritime safety in real-world port environments.

Power systems have been undergoing radical changes in recent years, and their planning and operation will be surely undertaken according to the Smart Grid (SG) vision in the near future. The SG initiatives aim at introducing new technologies and services in power systems, to make the electrical networks more reliable, efficient, secure and environmentally-friendly. In particular, it is expected that communication technologies, computational intelligence and distributed energy sources will be widely used for the whole power system in an integrated fashion. In particular, nowadays, unprecedented challenges like as stringent regulations, environmental concerns, growing demand for high quality, reliable electricity and rising customer expectations are forcing utilities to rethink about electricity generation and delivery from the bottom up. Moreover, the availability of low cost computing and telecommunications technologies, new generation options, and scalable, modular automation systems push utilities to be dynamic, innovative and ambitious enough to take advantage of them. Driven by the dynamics of the new energy environment, leading utilities, technology vendors and government organizations have created a vision of the next generation of energy delivery systems: the Smart Grid. Operational changes of the grid, caused by restructuring of the electric utility industry and electricity storage technology advancements, have created an opportunity for storage systems to provide unique services to the evolving grid. Especially Battery Energy Storage Systems (BESSs), thanks to the large number and variety of services they can provide, are powerful tools for the solution of some challenges that future grids will face. This consideration makes BESSs critical components of the future grids. The BESS can be applied for different services into the different levels of power system chain to satisfy technical challenges and provide financial benefits. In the context of the application of BESSs in SGs, there are two main problems that need to be addressed in a way that exploits the BESS potential, that are linked to their operation and sizing. This thesis focuses on both these aspects, proposing new strategies that allow optimizing the BESS adoption. When dealing with BESSs, sizing and operation are strictly linked. The correct sizing of a BESS, in fact, needs to take into account its operation which in turn will be effected with the aim of optimizing the whole system where it is included. In the first part of this research study, advanced optimal operating strategies were proposed for BESSs by considering both the distribution system operator perspective and the end user. Thus, the proposed operating strategies were performed with the aim of (i) leveling the active power requested by the loads connected to a distribution system (distribution system operator service), (ii) reducing the electricity costs sustained by an end-use costumer that provides demand response (DR) (end user service) and (iii) scheduling a microgrid (µG) with DR resources such as Plug-in Electric Vehicles (PEVs) and Data Centers (DCs) (both the two section service). The proposed strategies also satisfied technical constraints of BESSs and other components of the µG. The second part of the thesis presented the optimal sizing of BESSs aimed at maximizing the benefits related to their use. In the thesis, the sizing, which is performed by considering the end user point of view with reference to both the industrial and residential customers, is effected by adopting both deterministic and probabilistic approaches. With reference to the deterministic approach, a simple and quick closed form procedure for the sizing of BESSs in residential and industrial applications was proposed. In case of probabilistic approach, the case of a BESS installed in an industrial facility was considered and the sizing was performed based on the decision theory. Technical improvements and economic benefits of optimal operation and optimal sizing of BESSs in SG are demonstrated by the obtained results which are reported in the numerical applications. More specifically, it was clearly determined that BESSs can offer technical supports into the distribution operator section of the grid in terms of load management and security challenges. Moreover optimal integration of BESSs into the grid was also appealing for end users thanks to valuable amounts of electricity bill cost reduction. Regarding the original contribution of the thesis, the following considerations can be done. With reference to the load leveling service, an innovative two-step procedure (day-ahead scheduling and very short time predictive control) was proposed which optimally controls a BESS connected to a distribution substation in order to perform load leveling. In case of DR, a proper control of the BESS was proposed in order to perform DR under different price schemes, such as Real Time Pricing (RTP) and Time of Use (TOU) without modifying the daily work cycle of the industrial loads. The control procedure allows achieving contemporaneously two important goals that are the reduction of the bill costs and the prolonging the battery's lifetime so further reducing the costs sustained by the customer. With reference to the scheduling of microgrids, the original contribution of the thesis is focused on the proposal of optimization strategies aimed at managing and coordinating, simultaneously, batteries on board of vehicles or equipping data centers' Uninterruptable Power Supply (UPS) and Distributed Generation (DG) units. Also comparisons among different single-objective based strategies are made in order to highlight the most convenient. With reference to the sizing based on deterministic approach, unlike the other relating literature, the innovative contribution is that the closed form procedure takes into account both the technical constraints of the battery and contractual agreements between the customer and the utility. Moreover, in the economical analysis performed for the sizing, which is applied with reference to both residential and small industrial customers and is based on actual TOU tariffs, a wide sensitivity analysis to consider different perspectives in terms of life span and future costs was performed. Some aspects that affect the profitability of the battery, such as technological limitations (e.g. the battery and converter efficiency), economic barriers (e.g. capital cost and the rate of change of the cost) and variation of the load profile along the years were deeply analyzed. In case of sizing based on probabilistic approach, the original contributions of the thesis are mainly referred to the proposal of a new method that uses a decision theory-based process to obtain the best sizing alternative considering the various uncertainties affecting the sizing procedure. The thesis is organized in three chapters which are dealing with integration of BESSs in SGs. The first chapter reports basic concepts and characteristics of BESSs, fundamental components and features of SGs and different services that BESSs can provide. The optimal operation strategies of BESS are considered in second chapter which includes their problem formulation, solving procedures and results. The third chapter deals with the optimal sizing problem of BESSs for which the problem formulation, solving procedures and results are reported. Finally, the conclusions are presented in the last part of thesis.

Introduction. Environmental protection, rational use of natural resources and environmental security are among the guidelines for sustainable development of any state. However, it is important to emphasize that effective realization of such activities is possible only in peaceful time. At present, the full-scale Russian invasion of Ukraine puts under the direct threat effective ensuring of environmental security. The purpose of the paper is to analyze the environmental impact of the 2022 Russian invasion of Ukraine, highlight its consequences and predict the future environmental situation. Results. Under international law, an environmental crime is interpreted as an intentional attack when it is known that it will cause significant, long-term and serious damage to the environment, which is clearly disproportionate to the specific and directly expected overall military superiority. The military action of the Russian Federation in Ukraine in 2022 is undoubtedly an unjustified and unprovoked act of aggression against independent and sovereign Ukraine, as well as a violation of all existing norms of international law on environmental protection, sustainable development, humanitarian law, basic morals and principles of human coexistence. By May 2022 27 254 cases of environmental war crimes with regards to Russia's war against Ukraine had been registered. These include shelling and bombing of industrial and energy facilities, forest fires, sabotage of oil depots, pollution of the seas, destruction of reserves, as well as other horrific, inhumane actions of the Russian army. Conclusion. Russia's aggression against Ukraine has caused significant damage to the environment, like deterioration of soil, radiation pollution, biodiversity losses, forest fires, ecosystems’ destruction, etc. In order to remedy the current dire environmental situation, the development and implementation of urgent legal and policy measures is required in order to minimize the harmful environmental effects. In particular, it is necessary to constantly keep records of all crimes against the environment committed on the territory of Ukraine, and to record the damage caused. In addition, a special attention should be paid to the development of a plan for the reconstruction of Ukraine, which would include measures to restore and preserve the environment.

The use of photovoltaic systems connected to electrical grid for industrial, commercial, and residential facilities has been increasing. These systems require power converters to manage the energy converted by the photovoltaic panels. However, their operation depends on the presence of solar irradiance. In addition, the role of these converters has been limited to inject active power to the grid without considering power quality capabilities in electric facilities. In this work a double-loop control scheme is presented to facilitate this improvement. This scheme enables in the inverter the additional function of compensating reactive power from inductive-resistive loads to improve the power factor independently of the energy converted by photovoltaic panels. The system performance, with the proposed control algorithm, is verified by computational simulations using MATLAB/Simulink.

Increasing economic importance of the Arctic, further intensification of northern sea routes, and exceptional sensitivity of the arctic natural environment to anthropogenic impacts are fundamental factors for a comprehensive study of environmental aspects in the application of innovative technologies for the development of infrastructure in the Arctic. Despite the growing interest in low-power nuclear power plants as a distributed generation facility, their possible application in technologically isolated power systems does not lose relevance. The development of both the Arctic and Far Eastern regions of the Russian Federation presents great opportunities and demand for the use of nuclear power sources. Also, development programs for the Russian arctic zone imply a significant increase in the role and number of nuclear power facilities, in other words of potential radiation-hazardous facilities. Large-scale use of nuclear-powered installations in the Arctic necessitates advanced development of a scientifically grounded and modern forecasting system as well as assessments of threats and risks in case of possible radiation emergencies at nuclear- and radiation-hazardous facilities. Also, the development of proposals for necessary measures to minimize negative consequences of such emergencies is required. This is especially true for the case of compact placement of industrial, infrastructure and residential facilities in the Arctic in the immediate vicinity of nuclear facilities. The paper demonstrates that the demand for low-power nuclear power plants and their competitiveness will grow steadily in the conditions of electric-power industry decentralization, further spread of distributed generation and the development of technologically isolated power systems. Approaches to the generation of a low nuclear-power system based on the philosophy of industrialization of production and centralized management are presented. Special features of the environmental impact assessment of low-power nuclear power plants for the development of a methodology to study the radio-ecological hazard related problems are provided.

Nowadays, the use of heat pumps (HP) of air split-conditioners in air conditioning and heating systems of small industrial, public and residential facilities is becoming more common. It is known that the nominal heat capacity of HP of air split-conditioners is given in catalogs or reference literature under standard outdoor temperature conditions, namely: outdoor air temperature +7oС, indoor air temperature +21oС. At the same time, manufacturers of air split-conditioners do not ensure that, regardless of the size of heating capacity, all air split-conditioners have the same internal temperature conditions, namely: the evaporation temperature of the refrigerant and its condensation temperature. In this case, the thermodynamic efficiency, which can be best assessed by the exergetic output-input ratio (OIR) of different heating capacity of HP of air split-conditioners, is different; this, in our opinion, is incorrect. However, today there is a lack of mathematical models of bringing the operation of air split-conditioners HP to the similar internal temperature conditions, which will allow us to obtain the same exergetic OIR for different heating capacity of HP. To create the mathematical model of bringing the operation of HP of air split-conditioners to the equal internal temperature conditions, we have proposed them, namely: the evaporation temperature of the refrigerant +0.7oC and its condensation temperature +40oC. Taking these temperatures on the basis of the heat balances of the HP evaporator and HP condenser of air split-conditioners, we obtained the dependences for calculating air flow rates on the evaporator and condenser, which respectively maintain the proposed temperatures.

The paper investigates energy efficiency issues that arise in the operations of the fuel and energy complex within the institutions of the Penal System of the Russian Federation. It focuses on the main consumers of thermal energy, including heat supply systems for industrial and residential facilities. The problems of energy-efficient control of heating devices in the energy sector of the penitentiary department and the main ways of their solution are covered. The methods of designing energy-efficient controllers for increasing the efficiency of energy facilities within the Penitentiary System are considered. The issue of improving energy efficiency is presented in the form of an optimal control problem, for which an analytical solution is developed.