1 Introduction

The efficient use of resources needed for the sustainable development of modern industry cannot be achieved without using optimization methods and algorithms for engineering applications. Sustainable development assumes the interaction of various systems, such as goods production, energy, water, and environment, by optimally using waste from one as a resource in another (Mikulčić et al. 2017). As Mikulčić et al. (2021) also highlighted, “Global warming and climate change call for urgent minimization of the impact of human activities on the environment.” Consequently, energy, water and environment systems must integrate and become more sustainable. In line with that, this research field has drawn the attention of scientists, politicians and managers who recognize the urgent need for optimizing engineering solutions across various life-support systems.

Since 2002, Sustainable Development of Energy, Water and Environment Systems (SDEWES) Conferences have been regularly organized to advance the progress in this topical area. In 2022, three SDEWES events were held in hybrid and online modes:

  • On 22 – 26 May, the 5th South-East European SDEWES Conference (SEE SDEWES 2022) in Vlorë, Albania;

  • On 24 – 28 July, the 3rd Latin America SDEWES Conference (LA SDEWES 2022) in Sao Paulo, Brazil;

  • Finally, on 6 – 10 November, the 17th SDEWES Conference (SDEWES 2022) was held in Paphos, Cyprus.

Overall, SDEWES 2022 conferences have brought 817 researchers from 60 countries. The resulting set of 750 oral presentations constituted a forum to exchange and discuss their ideas and findings and identify challenges on regional, national, and international scales. The research topics ranged from technical, economic, and social studies to the research works delving into the integration and sustainability of goods production, energy, transport, water, and environment systems. A decision-making problem can often be translated into a mathematical model and solved using suitable methods and computational tools. This Special Issue (SI) collected contributions dealing with recent advances in optimization methods and algorithms to integrate life-supporting systems. Ten papers presented at the 2022 SDEWES Conferences constitute the body of the current SI of the Optimization and Engineering (OPTE) journal.

2 Background

This section’s content builds on papers published in SIs of other journals dedicated to the SDEWES conference series. It provides the background to Sect. 3, where papers included in this SI are reviewed.

Complimentary to the extensive network communications and exchange of ideas, SDEWES events involve a comprehensive publishing strategy. This SI is the fourth in collaboration with the OPTE journal, in line with the systematic collaboration with other high-quality journals. It expands the knowledge body created by the previous OPTE SIs: the first one, dated January 2021 (Mikulčić et al. 2021); the second SI, dated July 2021 (Trafczynski et al. 2021) and the most recent one, dated September 2022 (Trafczynski et al. 2022).

This section briefly summarizes the papers covering diversified research fields, including energy and water issues, environmental engineering and management, and planning for sustainable development. The papers were published in other journals’ SDEWES-related SIs over the last few years. Many have solved complicated data collection and decision-making challenges using sophisticated mathematical modelling and numerical techniques.

The paper by Costa et al. (2020) considered achieving high efficiency and low pollutant emissions of a syngas-powered engine through control of the spark timing and the air-to-fuel ratio, with optional exhaust gas recirculation. The authors addressed this issue by applying model-based design and multi-objective optimization The aim was to enhance power generation's energetic and environmental performances under flexible fuel quality. The identified optimal solution made it possible to reduce both nitrogen oxides and carbon monoxide emissions by up to 50% while avoiding the reduction of power output.

Hybrid electric power trains are often applied in road vehicles to improve fuel economy and reduce greenhouse gas (GHG) emissions. Minimizing fuel consumption requires finding optimal control strategies to match different power-train operating regimes. Cipek et al. (2020) discussed the advantages of hybrid electric vehicle control based on a cascade approach to control variable optimization. Using dynamic programming and a gradient-based optimization algorithm, the authors significantly reduced the computational effort and increased the precision of the globally optimal result.

Increasing applications of air pollution monitoring contribute to growing awareness of pollution effects on human health. Heydari et al. (2022) proposed a new hybrid intelligent model based on long short-term memory (LSTM) and a multi-verse optimization algorithm (MVO) to enable prediction and analysis of the air pollution generated by Combined Cycle Power Plants. He developed model uses the LSTM component to forecast the emissions of nitrogen and sulfur oxides produced by the plant. At the same time, the MVO algorithm optimizes the LSTM parameters for reducing the forecasting error.

Petrucci et al. (2022) proposed an innovative algorithm for community energy management control. It assumes that customers are involved in energy trading and take advantage of their energy flexibility. A matrix-based predictive control system uses the strategy of simultaneously considering individual and community preferences. The authors presented a concept of system structure involving control volumes that applies to energy communities differing in size, energy carriers, penetration of photovoltaics (PV), and number of electric vehicles. The system relies on a neural network for real-time prediction of buildings’ energy demand and generation of data for energy flow optimization.

Allen et al. (2022) presented a topology optimization approach for district heating systems. A subset of buildings is selected using particle swarm optimization, and the network topology for any candidate subset of buildings is optimized employing graph-theory-based heuristics. The research results indicate thattopology optimization can significantly reduce life cycle cost, source energy use intensity, and CO2 emissions.

Energy efficiency is a critical factor affecting the profitability of industrial process systems. While heat exchanger network (HEN) retrofit is known as a powerful tool for energy-saving and heat Integration, the retrofit cost is drastically increased if topology modifications with new heat exchangers and re-piping are required. Li et al. (2021) investigated ways to avoid network topology modification when heat transfer enhancement is applied as part of the HEN retrofit. The authors developed a target-evaluation method for optimal HEN retrofit, considering the costs of improving thermal efficiency and enhancing heat transfer.

Calise et al. (2022) presented the design and optimization of a 5th generation district heating and cooling network for a residential district on the Mediterranean island of Pantelleria.The aim was to evaluate the attainable energy, economic and environmental savings. The focus was on a novel network arrangement employing two low-temperature neutral rings with fluid-temperature control through two groups of seawater-cooled heat pumps. The studied arrangement also includes wind turbines and PV panels for power generation. The dynamic simulation model, thermo-economic analysis, and environmental study of a 5GDHC network—which connects to a renewable energy plant with a solar field and WTs—are presented in this paper. The network employs saltwater as a thermal source. The TRNSYS programme was used to run the simulations. The primary component of the architecture under investigation is a network built around two water loops, the RING1 and RING2, which run at 16–19 °C and 18–25 °C, respectively.

In the study by Solis et al. (2021), a multi-objective decision-support model was used for biorefinery optimization. Simultaneously, minimizing the cost and environmental impact was performed, adopting the resource recovery and recirculation principle and employing the LCA methodology to resolve the system's environmental impacts. An algal biorefinery producing biodiesel, glycerol, biochar, and fertilizer was considered in a case study. It was found that the optimal results strongly depend on demand fluctuations and process unit efficiencies.

Combined heat and power (CHP) generation plants are widely recognized as valuable solutions to reduce primary energy consumption and carbon dioxide emissions. These solutions’ primary energy saving and CO2 reduction potentials require accurately defining and managing heat and electricity loads. Bartolucci et al. (2022) applied a two-level optimization approach for designing a CHP system in a hospital building based on historical data on the demand for heat and electricity. First, using clustering analysis, a set of representative annual load patterns was identified for use as input data in the plant design. Second, Mixed Integer Linear Programming (MILP) and a Genetic Algorithm were used to optimize the energy dispatch and CHP plant output for maximum primary energy saving and minimum total costs and CO2 emissions.

Kim et al. (2020) studied the optimal operation of a microgrid in which the generating capacity can be committed to the charging/discharging of energy storage systems (ESSs). They developed a model featuring an inter-temporally constrained mixed-integer nonlinear programming (MINLP) problem and selected a parallel computation method based on problem decomposition for optimizing the solution. The proposed approach was tested on the CIGRE medium-voltage microgrid benchmark system. The simulation results confirmed the suitability of the mathematical model and optimization method for application to the real-time operation of the microgrid system.

Alirahmi et al. (2022) studied the co-generation of power, hydrogen, oxygen, and hot water in an innovative solar-driven energy system that uses parabolic trough collectors (PTCs) for supplying heat to cascadedsteam and organic Rankine power cycles (SRC and ORC). The system also includes a thermoelectric generator (TEG) in which heat discharged from the ORC condenser is transformed into additional electricity for the electrolyzer. A multi-objective approach was adopted to minimize the specific cost of the system output and maximize exergy efficiency simultaneously. The findings also indicate that using a TEG enhances efficiency and decreases the specific cost of the system output. This research introduces and examines a novel solar-powered energy system capable of generating multiple forms of energy. The key components of this system consist of PTCs, SRC, ORC, and an electrolyzer. Additionally, the study investigates the possibility of using TEG for electricity generation to replace the ORC condenser. Furthermore, the system's optimal operating conditions are determined through multi-objective optimization using a genetic algorithm. To ensure a realistic interpretation of the results, the performance of the proposed system is evaluated in four cities in the Khuzestan province of Iran.

The work by Lagouir et al. (2022) proposed an innovative daily energy management system for dispatch optimization and operation control of a typical microgrid power system. The authors formulated the dispatch problem as a simultaneous minimization of the operating cost, pollutant emissions and power conversion losses. Pareto optimal solutions are generated using the weighted sum method. Then, the best compromise solution is found by employing the fuzzy set approach and the ant-lion optimization method for performing the computational task.

Significant energy savings can be achieved in many commercial buildings that also exhibit potential for the application of renewable technologies. Roumi et al. (2022) investigated the energy consumption in a reptile exhibition building. They calculated the heating and cooling demand and the electricity demand, including lighting, heating, cooling, ventilation, etc. Various scenarios of combined wind and solar energy applications were studied, taking economic and environmental aspects into account. The finding was that wind energy has the lowest levelized cost of electricity; however, it cannot cover the electricity demand for the reptile garden. In HOMER software, the system is modelled. The initial goal of this study is to determine the best mix of renewable resources at the lowest possible cost. The ideal number of PV panels, converters, and micro wind turbines for the case study are all established. HOMER determines the amount of power generated in each scenario based on the particular geographic location of the case study and the quantity of devices utilised. The grid supplies the shortfall in power, while surplus electricity is sold to the grid.

The reduction of CO2 emissions is widely recognized as a challenging issue. In particular, the high energy demand for CO2 capture from combustion processes decreases power generation efficiency. Zach et al. (2022) studied membrane-based post-combustion capture aiming to minimize energy demand. The authors used the network flow approach to develop a CO2 capture system mathematical model and employed external simulation modules for nonlinear problems to optimize the solution. A case study was performed to verify the model’s suitability for process optimization.

CO2 capture processes employ chemical absorption that requires the use of absorbents such as amines and is characterized by a high energy demand. The research done by Wang et al. (2022) was aimed at optimizing the amine solution mixing ratio to increase the absorption capacity and cut down the energy consumption. The authors proposed a derivative analysis of the standardized vs variables diagram (DSVD method) as a graphical tool to ascertain the optimal mixing ratio that maximizes the benefit and minimizes energy demand.

Herc et al. (2022) optimized the intermediate steps towards energy system decarbonization. The objective was to attain predetermined levels of renewable energy and emissions at minimum system cost and limited use of natural resources. Energy-generating capacities, demand response technologies, and energy storage were considered as decision variables. The authors used software tools EnergyPLAN (for energy planning) and EPLANopt (for Python-based optimization). The research highlighted the importance of systematically implementing flexible generating capacities and demand-response approaches, such as vehicle-to-grid.

Delivering low-carbon heat in an energy system requires using electricity and hydrogen as low-carbon alternatives to natural gas. Aunedi et al. (2022) connected two advanced investment-optimizing models, RTN (Resource-Technology Network) and WeSIM (Whole-electricity System Investment Model), to evaluate cost-efficient heat decarbonization pathways for the UK. According to the research results, the hydrogen from gas reforming with CCS can be applied in the medium term. In the longer term, hydrogen is a viable option for supplying peak heat demand, with electric heat pumps covering the bulk of heat demand.

The articles reviewed above are only examples of papers from recent SDEWES conferences on optimizing energy, water and environment-related developments. At the same time, these articles confirm the need for further progress within the reviewed research fields.

3 Overview of the current special issue topics and papers

Out of the presentations at the SDEWES 2022 conferences, sixteen contributions were selected as candidates for the SI of the OPTE journal. After being reviewed by SDEWES International Scientific Committee members, SDEWES Scientific Advisory Board members, and external experts, ten papers were approved and included in this SI. These papers cover topics related to data collection and decision support in engineering problems across all SDEWES fields.

Kodba et al. (2023) delved into the potential of anaerobic systems for biogas production, integrating renewable energy generation, waste management and treatment, and fertilizer production. The authors presented an approach emphasizing the economic optimization of a biomass supply network for biogas production in urban areas considering availability of biowaste and residues in restaurants, shops, and the food and drink industry. Firstly, the model employed the Geographic Information System (GIS) to introduce a GIS-based method that sets a maximum allowable transport distance, aiming for greenhouse gas(GHG) savings from biogas usage. It aligns with Directive 2018/2001, which mandates GHG emissions from biogas production reduced by 80% until 2026. Secondly, the study relies on GIS mapping of biomass resources and P-graph concepts to optimize the biomass supplies. The model, developed using P-Graph Studio and QGIS tools, is tested assuming two levels of annual production: 36,000 GJ and 72,000 GJ. The results reveal the most profitable solutions for both cases and recommend concentrating biogas production in a single brewery. The specific costs, including feedstock and transport, are also evaluated.

In pursuing sustainable energy solutions, hydrogen emerges as a pivotal player in the global energy landscape. Lampe and Menz (2023) studied the potential of generating renewable hydrogen through water splitting via thermochemical two-step redox cycles with energy supply from concentrated solar power. This method showcases a superior solar-to-fuel efficiency when juxtaposed with the combination of PV and subsequent electrolysis. The research sheds light on optimizing the operational process strategy of a 50 kW pilot plant. This plant employs a fixed bed reactor operated in a temperature-swing regime with ceria as the reactive material. The study used a physical system model, previously validated by two measurement campaigns, to evaluate key process parameters, including operation temperatures, fluid flows, and switching times. MATLAB surrogate optimization algorithm and neural network surrogate optimization were employed in solution finding. Assuming reduction temperatures between 1400 and 1700 °C, the maximum efficiency is estimated at up to 5.6%. The study also touches upon the economic implications of the technology. In essence, the research underscores the potential of solar-driven thermochemical processes in optimizing hydrogen production and emphasizes the role of operational strategies in enhancing efficiency and sustainability.

Recently observed rapid growth of data centres (DCs) has led to increased electricity consumption. A significant portion of the energy consumed in DCs is transformed into heat, elevating the operating temperature of DC components, and is typically discarded without further use. Recognizing the potential of this waste heat, researchers have explored its integration into district heating (DH) systems as a means to reduce reliance on conventional heat fuels. Miškić et al. (2023) examined three methods of waste heat integration: using a heat exchanger (HEX), a supporting heat pump, or a combination of the two. The authors employ a combination of a heat pump and HEX in their study. Despite existing implementations of these methods, there remains a gap in research on optimizing waste heat integration into the district heating network (DHN). To address this, the authors developed models of thermodynamic processes in the DC and complementing pinch analysis (PA). Their research introduces a method to evaluate the economic feasibility and optimize the DC waste heat integration technology into DH systems using an hourly merit order based on PA. The methodology's effectiveness is illustrated with a case study of a data centre in Zagreb, Croatia. The findings confirmed that waste heat from data centres is a valuable resource for district heating, offering both environmental and economic benefits.

Rosecký et al. (2023) addressed the challenges faced in logistics and infrastructure planning modelled as the network flow problem. While insightful, these tools face computational challenges when solving large-scale, detailed problems. The paper introduces a classification approach aimed at reducing the number of variables. Firstly, using the design of experiments, solvable smaller problems are created using artificial data. The probability of each arc's presence in the optimal solution is then estimated considering cost and capacity aspects. After training on selected problems, the model is applicable to others. A practical application of this framework is demonstrated through waste management planning in Czechia. The authors showed that using only 5% of arcs is sufficient to analyze the objective function. This reduction leads to a computing time of only 7% of the original task. The study underscores the potential of machine learning in optimizing network flow problems, offering a significant reduction in computational demand without compromising the quality of solutions.

Kůdela et al. (2023) analyzed combined heat and power stations as pivotal components of district heating systems. These stations, equipped with multiple boilers, harness fossil fuels and renewable resources to produce steam for heating and electricity generation. The study emphasized the importance of maintaining boiler operations either continuously above a specified threshold or switching them off to optimize efficiency. The authors introduced a two-tiered approach for the optimal control of these stations over a set time frame. The foundational level focuses on determining the best operational parameters for the station, considering hourly steam demands. In contrast, the overarching level strategizes the optimal scheduling of individual boiler operations throughout the time considered. The lower echelon is articulated as a MILP challenge, tackled using parametric programming, while the upper stratum employs a dynamic programming algorithm with a rolling horizon. A comprehensive case study validates the method, exploring its computational complexity across varied parameters, steam demands, and electricity price fluctuations. This approach is claimed to be a blueprint for addressing similar control challenges in other systems with similar structures.

Li et al. (2023) explored the intricacies of microchannel heat exchangers, known for their high heat transfer efficiency and compactness. The paper emphasized the significance of accurate microchannel wall temperature measurements, which directly influence the heat transfer coefficient's measurement results. Despite the water bath's ability to effectively control the heat flux in microchannel experiments, there lacks a robust method for a precise measurement of the microchannel's wall temperature. Employing the CFD method for fluid–solid coupled heat transfer, the authors scrutinized the measurement errors of six wall temperature measurement schemes within a Reynolds number range of 2400–7200. The study delved into these schemes' flow characteristics, introducing an optimized method to measure the microchannel wall surface temperature under a flowing water bath using a standard resistance temperature detector. The findings reveal that the error of wall temperature measurement due to the CFD method can be minimized to 0.0137 K. Overall, this research offers valuable insights into enhancing the accuracy of wall temperature measurements in microchannel heat exchangers.

Pavković et al. (2023) studied automatic tuning and estimation of key parameters of the drill-string system for deep drilling. The focus is tuning an external vibration suppression system, which employs a speed control loop of the drilling electrical drive. This system, uses drill-string torque feedback for torsional vibration damping. The tuning procedure is activated when the drill string is lifted from the well's bottom, and the drilling electrical drive enters an oscillatory behavior. The auto-tuning system incorporates a process model of the torsional oscillator and a phase-locked loop. The control system’s effectiveness and the robustness of the drill-string parameter estimator have been validated through comprehensive simulations using field data.

Xie et al. (2023) studied dynamic water injection demands for oilfields across various development phases. The study emphasized the importance of matching the characteristics (layout and size) of the water injection pipeline network to fluctuating demands and advocated for long-term planning to circumvent frequent network expansions. The research introduced a MINLP model for optimizing the characteristics of the water injection network, specifically for network retrofit. The model's objective is to minimize both construction and operational costs. It incorporates pipeline and truck transport, factoring in constraints related to water flow characteristics and supply–demand balance. The model's suitability was demonstrated using a case study from an oilfield in China, focusing on the future water injection planning of an existing pipeline network across three development periods. The findings revealed a cost-saving of 14.6% in the final period compared to the initial period, primarily attributed to the decrease in truck-transported water volume.

The increasing electrification of residential heating and cooling systems has led to a surge in the use of electrically driven heat pumps combined with thermal/electrical energy storage. The challenge lies in devising intelligent control mechanisms for these systems to maximize the share of renewable energy. Diller et al. (2023) introduced a novel approach leveraging dynamic programming (DP) as a method for smart controls. This method is particularly promising due to its resilience against the complexities associated with renewable energy sources. The authors presented a modelling approach that employs reduced order models (ROM) of the main components of a heat pump substation. Through this approach, they demonstrated the potential for significant cost savings, highlighting power consumption 13% lower compared to traditional rule-based control methods. The results underscore the efficacy of the DP-based method in optimizing the operation of heat pump systems with thermal energy storage. By integrating renewable energy sources and employing smart control strategies, the proposed method offers a sustainable solution for the residential heating and cooling sector, paving the way for a more energy-efficient future for such systems.

Pursiheimo et al. (2023) introduced “Predicer” (Predictive Decider), an open-source modelling package designed for multi-market day-ahead energy and reserve market operations. Predicer determines decision variables and bid matrices using scenario-based stochastic optimization. Its primary objective is to maximize the expected risk-adjusted profit value over a specified timeframe. The energy system structure within Predicer is abstract and allows users to model diverse energy systems and establish connections between assets, commodities, energy markets, and reserve markets. It is possible to account for various properties, including unit ramp rates, online units, dynamic energy storages, market realization, and market bidding requirements. Energy and reserve opportunities can be aggregated into a virtual power plant, allowing asset owners to develop optimized portfolio-level bids for different market products. A user can define model scenarios by time-series forecasting for such factors as market prices, renewable energy supply, energy demand, etc. Predicer design aims to support energy and reserve market operations for systems of varying scales and facilitate the creation of bidding curves based on stochastic optimization.

4 Conclusions

The main part of this article builds on ten recent research papers on optimizing engineering solutions to facilitate the sustainable development of Energy, Water and Environmental Systems. Recent advancements in optimization and engineering have directed significant attention towards developing and enhancing processes that contribute to sustainable and efficient energy systems. A recurring theme across the SI articles is the exploration of optimization techniques in various applications, from energy systems to industrial processes. Several articles delve into optimizing renewable energy systems, with emphasis on the potential of hydrogen production. The integration of machine learning techniques, especially neural networks, into optimization processes has been evident in the research. These techniques aid in pattern recognition, forecasting, and enhancing the efficiency of optimization algorithms. They play a pivotal role in predicting system behaviors, thus facilitating better decision-making in real-time scenarios. Operational strategies, heat transfer mechanisms, and reactor designs for these processes have been optimized to maximize output and minimize energy consumption. Beyond the technical aspects, economic considerations are crucial in optimizing the solutions. The viability of implementing solutions in real-world scenarios is contingent on their economic feasibility. As such, several articles have incorporated economic analyses to ensure that the optimized solutions are technically sound and economically viable. The optimization processes discussed multiple disciplines, integrating knowledge from thermodynamics, chemical engineering, and computational modelling. This multidisciplinary approach ensures a holistic understanding of the problems at hand and facilitates the development of comprehensive solutions. The research points towards an increasing emphasis on sustainable energy solutions, with renewable energy at the forefront. The continuous refinement of optimization techniques and technological advancements promise a future where energy systems are more efficient, sustainable, and economically viable.

Developed by author teams representing more than a dozen countries, the reviewedcontributions were selected to illustrate international research efforts in the said topical area. The editors of this fourth Special Issue of the OPTE journal believe that SI content will attract the interest of its readers.