What Can We Do in Bucharest? The Issues of Decarbonising Large District Heating Systems

Abstract

District heating systems are central to Europe’s decarbonisation strategy and its 2050 climate-neutrality objective. However, district heating is deeply embedded in the socio-economic system and the built environment. This makes compliance with policy targets at the local level particularly challenging. The issues are attributable to two factors. Firstly, the process is characterised by a high degree of complexity and multidimensionality. Secondly, there is a scarcity of local resources (e.g., land, surface waters, waste heat, etc.). In Bucharest, Romania, the largest district heating system in the European Union, the process of decarbonisation represents a particularly complex challenge. The system is characterised by large physical dimensions, high technical wear, heavy dependence on natural gas, significant heat losses and complex governance structures. This paper presents a strategic planning exercise for aligning the Bucharest system with the Energy Efficiency Directive 2023/1791. Drawing on system data, investment modelling, and local resource mapping from the LIFE22-CET-SET_HEAT project, the study evaluates scenarios for 2028 and 2035 that shift heat generation from natural gas to renewable, waste heat, and high-efficiency sources. The central objective is the identification of opportunities and issues. Options include large-scale heat pumps, waste-to-energy, geothermal and solar heat. Heat demand profiles and electricity price dynamics are used to evaluate economic feasibility and operational flexibility. The findings show that the decarbonisation heat supply in Bucharest is technically possible, but financial viability hinges on phased investments, interinstitutional coordination, regulatory reforms and access to EU funding. The study concludes with recommendations for staged implementation, coordinated governance and socio-economic measures to safeguard heat affordability and system reliability.

1. Introduction

Existing European district heating systems are undergoing a profound transformation to meet the EU’s 2050 decarbonisation goals. The revised Energy Efficiency Directive (EED) [1] establishes that by 2028, efficient district heating (DH) systems must supply into the district heating network (DHN) at least 5% of heat from renewable sources (RES). By 2035, 35% of heat supplied to DHN must be from either RES or waste heat sources (WHS), and in total RES, WHS, and high-efficiency cogeneration (CHP) must contribute with at least 80%. The respective thresholds are set for 2040 and 2045, and by 2050, an efficient district heating system (EDHS) must use only renewable energy, only waste heat, or only a combination of renewable energy and waste heat.

The focal point of this study is the strategic planning of the energy transition and decarbonisation process for large DH systems. A growing body of literature is currently investigating technological and strategic solutions for such systems. Lund et al. [2,3] introduced the fourth-generation district heating (4GDH) concept, promoting lower temperature operation and integration with renewable electricity. Volkova et al. [4] and Sihvonen et al. [5] highlighted large-scale heat pumps and sector coupling as key enablers of future heating systems. Pakere et al. [6] demonstrated that multi-source DH systems can achieve full decarbonisation by combining solar, biomass, and waste heat supported by seasonal storage. Pettersson et al. [7] and Pursiheimo et al. [8] emphasised the importance of integrated planning and compared heat pump and nuclear options for Nordic cities. Kubin et al. [9] presented DH transformation planning results, which confirmed that flexible sector coupling technologies form the backbone of the most economically efficient transition pathway. They highlighted the risks of rapid scale-up, especially regarding permitting, investment, and supply chain conditions.

Dzierzgowski and Cenian [10] proposed a decarbonisation methodology that focuses on substation modernisation and regulation, followed by the implementation of low-carbon heat sources (including waste heat) and the implementation of digital measures (monitoring and modelling). A large-scale experiment performed in the DH system in Łomża, Poland, confirmed that the implemented methodology allowed a decrease in network supply temperature (from 121 to 96 °C) in the whole system, leading to lower heat losses, in turn reducing coal demand and CO₂ emissions (up to 30%), and an increase in heat transfer efficiency in buildings, grid hydraulic stability and improvement in the working condition of mixing systems in the CHP and heating plant.

Kalina et al. [11] reveal that in such systems, decarbonisation is technically attainable, yet far more complex than often assumed. The transformation faces substantial technical barriers. The high operating temperatures of existing networks limit the performance of heat pumps during the heating season, and the large-scale seasonal storage required to balance intermittent sources exceeds what is realistically possible. Even when technically feasible, the transition of DH systems introduces significant economic burdens. Under the study’s financial modelling, the overall investment package leads to a negative net present value unless supported by substantial external funding. The study concludes that decarbonisation cannot be treated as a narrow technological upgrade. Instead, success depends on broader structural changes, including lowering network temperatures, improving building efficiency, enabling sector coupling, adjusting legal frameworks, and establishing new business models for cooperation with waste heat suppliers.

The strategies announced by cities, such as Berlin, Amsterdam, Copenhagen, and Vienna, reveal that depending on local conditions, the future primary energy mix is likely to be highly diversified, resulting in increasingly complex DH systems. Although these strategies differ in pace, resource availability and governance, a clear pattern emerges. Each of the considered cities envisions a future district heating system built on low-temperature networks, large-scale heat pumps, extensive waste-heat recovery and the systematic integration of renewable electricity.

The strategy employed by Berlin is indicative of two interrelated factors. Firstly, it is a reflection of the legacy of a substantial, historically fossil-fuelled heating system. Secondly, it is a manifestation of the city’s growing political ambition [12]. Coal has already been phased out, and future supply is expected to rely heavily on industrial and data-centre waste heat, combined with large heat pumps powered by renewable electricity. Detailed Berlin’s heat planning process relies on heat maps and zoning tools to guide the expansion of district heating, the transition of buildings to individual heat pumps, and the gradual transformation of existing networks. Lowering supply temperatures is essential to integrate low-temperature resources, while bans on new fossil installations and updated efficiency standards reinforce the shift at the regulatory level.

Amsterdam approaches decarbonisation from a different angle, which is a binding political decision to eliminate natural-gas use in buildings by 2040 [13]. Publicly released heat transition visions outline which districts will move to DH, all-electric solutions or hybrid systems. Amsterdam’s plans emphasise rapid expansion of both high- and low-temperature networks and make unusually strong use of aquathermal energy (i.e., heat extracted from canals, rivers and wastewater) as well as large heat pumps and substantial geothermal potential where geological conditions allow. Waste heat from waste-to-energy plants and data centres serves as a further backbone of the system. The city pairs this technical rollout with a socially oriented governance framework that ensures affordability and transparency as gas networks are dismantled neighbourhood by neighbourhood.

Copenhagen, long considered a frontrunner in district heating, has articulated some of Europe’s most detailed long-term strategies. The city’s carbon-neutrality goal and the broader metropolitan “Heat Plan Copenhagen” [14] provide a system-wide blueprint that extends to mid-century. The plan foresees a marked reduction in network temperatures, enabling large heat pumps, geothermal installations and significant amounts of industrial and municipal waste heat to feed into the system. Copenhagen emphasises system optimisation through sector coupling, where district heating is positioned as a flexible component in a wider energy system dominated by wind and solar power. Large-scale thermal storage, including pit thermal energy storage, ensures that excess renewable electricity can be converted into heat and stored for later use. Remaining fossil or biomass-based CHP plants are expected to be gradually replaced or supplemented by carbon-neutral fuels and carbon-capture technologies. Power-to-X technologies and CO₂ capture installations are also taken into consideration.

Vienna’s trajectory is influenced by its 2040 fossil-free heating target, one of the earliest and most ambitious in Europe. The city has developed a heat-planning approach that distinguishes areas best served by district heating from those suited to individual heat pumps [15,16]. As part of this strategy, district heating is expected to supply more than half of the city’s heat demand by 2040, and the DH system itself must reach full climate neutrality by the same year. Large heat pumps powered by wastewater, river water and industrial waste heat form a central pillar. Deep geothermal energy, combined with heat pumps, is projected to deliver a substantial share of Vienna’s future district heating production, around 4 TWh annually, allowing the gradual retirement of fossil-fuelled CHP plants. At the same time, existing waste-to-energy facilities are to be progressively decarbonised through more efficient operation and potentially carbon-capture technologies. Vienna complements these technological measures with broad renovation programmes and a citywide initiative “Raus aus Gas”, which pilots alternatives to gas in municipal and social housing, ensuring that building-level interventions align with the transformation of the heat networks.

Malcher et al. [17] evaluated how different decarbonisation strategies reduce emissions in national district heating systems by applying an integrated energy-and-emissions model combined with a derivative-based sensitivity analysis. Focusing on Sweden, France, Germany and Poland, they examined widely discussed measures, such as increasing low-carbon heat sources, lowering supply temperatures, improving efficiency, deploying power-to-heat technologies, and modifying the role of CHP. Results show that the effectiveness of these strategies varies strongly by each country’s energy mix and the extent to which CHP plants rely on fossil fuels. Expanding low-carbon heat sources is consistently the most impactful option. A 1% increase in their share can reduce emissions by 0.8–1.3 kg CO₂e/GJ of heat. In Sweden and France, where electricity is already largely low-carbon, heat pumps and electric boilers emerge as the next most effective measures. Conversely, in Germany and Poland, characterised by higher-carbon power, reducing heat demand and distribution losses delivers comparable emission cuts (0.7–1.3 kg CO₂e/GJ). Substantially lowering fossil-fuel-based CHP generation also achieves significant reductions, challenging policies that incentivise gas-fired CHP. Other options, including green hydrogen and CCS, offer noticeably smaller benefits.

The present study examines the feasibility of the transition and the associated trade-offs in an extreme-scale legacy district heating (DH) system, where developing suitable transition pathways involves significant technical, organisational, and economic challenges. The primary objective of this article is to draw readers’ attention to the limited degree of freedom, serious constraints, system balancing and production units dispatch problem, demand for large heat storage capacity, and poor profitability without subsidies or an increase in operating income. The study does not offer ready-made solutions; instead, it identifies opportunities and key issues, thereby paving the way for dialogue among relevant stakeholders.

The case study addresses Bucharest, Romania, where the DH system is the largest in the European Union. According to the IEA DHC definition [18], this is a second-generation system with a water network operating forward temperature ranging from 80 to 110 °C. The district heating network (DHN) has a total length of 3847 km and delivers heat to more than 500,000 individual households, 3611 industrial and 1122 other consumers. The current peak demand for heat approaches 2.0 GW, and the annual amount of heat delivered to the network is around 4500 GWh/year. The heating network covers around 72% of the heating demand of the Municipality of Bucharest. The remaining 28% is heat produced in various individual heating units.

The input energy mix is not diversified, and the system is almost entirely based on the combustion of natural gas. As for 2024, it experiences severe losses and frequent breakdowns. The key system components, i.e., production plants and networks, are worn out and in bad technical condition. Since the system has been underinvested for many years, the scale of the required renovation is enormous. A potential action plan is presented in this study, which examines possibilities to integrate different technological solutions.

2. Materials and Methods

This study was conducted within the framework of the LIFE Programme–co-funded project LIFE22-CET-SET_HEAT. It builds on analyses carried out during the project’s joint development and design activities between 1 October 2023 and 14 March 2025, including a series of co-creation workshops. During the workshops, project partners and invited guests jointly identified technological solutions and other measures to help the targeted DH systems in Bucharest (Romania), Opole (Poland), Vilnius (Lithuania) and Zagreb (Croatia), qualify as EDHS under the revised EED [1].

The objective of this study is to evaluate technically feasible pathways for decarbonising Bucharest’s large fossil-fuel-fired district heating system in accordance with the criteria for EDHS specified in the revised EED [1]. The specific goal is to meet the criteria of EDHS from 1 January 2028, by ensuring that the share of heat from RES in the total amount of heat supplied into DHN is at least 5% and the total share of RES, WHS or CHP is at least 50%. From 1 January 2035, the target is to ensure a minimum of 35% RES and WHS, and a minimum of 80% total share of these sources and CHP. Accordingly, the strategic planning process is subject to the following constraints:

$$u_{RES,2028}={\displaystyle \frac{{\sum }_t\sum _i{Q}_{i,\mathrm{R}\mathrm{E}\mathrm{S},t}}{{\sum }_t\sum _i{Q}_{i,t}}}\ge 0.05 \cup u_{RES+WHS+CHP,2028}={\displaystyle \frac{{\sum }_t\sum _i{Q}_{i,\mathrm{R}\mathrm{E}\mathrm{S}+\mathrm{C}\mathrm{H}\mathrm{P},t}}{{\sum }_t\sum _i{Q}_{i,t}}}\ge 0.50$$

(1)

$$u_{RES+WHS,2035}={\displaystyle \frac{{\sum }_t\sum _i{Q}_{i,\mathrm{R}\mathrm{E}\mathrm{S}+\mathrm{W}\mathrm{H}\mathrm{S},t}}{{\sum }_t\sum _i{Q}_{i,t}}}=0.35 \cup u_{RES+WHS+CHP,2035}={\displaystyle \frac{{\sum }_t\sum _i{Q}_{i,\mathrm{R}\mathrm{E}\mathrm{S}+\mathrm{W}\mathrm{H}\mathrm{S}+\mathrm{C}\mathrm{H}\mathrm{P},t}}{{\sum }_t\sum _i{Q}_{i,t}}}\ge 0.80$$

(2)

where u is the share of heat from particular types of heat sources in the total heat delivered to DHN, Q_i is the heat delivered from specific source i at hour t.

The research and planning methodology integrated making inventories of existing infrastructure, system diagnostics, system and energy mapping, scenario development, investment modelling and assessment, and hourly system simulations based on actual weather, load and market data. Reference [11] discusses the adopted methodological approach in detail. In the case of Bucharest, however, greater attention was paid to spatial planning and the location of individual heat sources in the dense urban area. Several web tools and public data sources [19,20,21,22,23] supported data collection, while the energyPRO™ version 4.9 software [24] was employed for system simulations. The operation of particular production units was optimised using the Mixed Integer Linear Programming Technique (MILP) and the open source HiGHS solver developed at the University of Edinburgh [25], and integrated in the energyPRO™. Costing and financial analyses were conducted using Microsoft Excel spreadsheet software.

The Bucharest DH network comprises 6 heating areas, so-called sectors, and 8 main production plants, namely Sud, Progresul, Grozăvești, Vest, Vest Energo, and Grivița, Titan and Casa Presei, with a combined installed heating capacity exceeding 1.73 GW_th. Forward temperatures range from 80 to 110 °C, with returns of 40–55 °C, characteristic of second-generation DH systems. The system boundary and location of heating plants are depicted in Figure 1. In addition, there are more than 46 local gas boiler plants of the total installed capacity of 257 MW, which are not delivering heat to the DH network but directly to buildings. The total heat demand density is depicted in Figure 2.

The DHN, which supplies heat to approximately 8200 apartment blocks, 320 buildings, and around 4900 institutions, is operated by Compania Municipală Termoenergetica București S.A. (CMTEB). It has a total length of 3847 km, of which 884 km form the so-called primary transmission network, and the remaining 2963 km form the secondary distribution network, to which buildings are connected. An important component of the network infrastructure is around 1027 so-called Thermal Points, which are group heat exchanger stations between primary and secondary networks. The location of these stations around the city is depicted in Figure 3.

Around 92% of DHN input heat is purchased from external suppliers. In total, depending on the year, approximately 74–80% of purchased heat is produced in steam cycle cogeneration units. CMTEB’s in-house heat production relies entirely on boilers. Figure 4 depicts an annual DHN heat supply profile, and Figure 5 shows the primary network regulation curves. Figure 4 also demonstrates an indicative level of required newly installed heating capacity to satisfy the EED requirements for EDHS.

The existing heating plants were put into operation between 1955 and 1982. Around 65% of the primary network and 46% of the secondary network are pipelines older than 25 years. The biggest problem with the network is the high corrosion of the pipes. Many parts of the network are damaged, causing major operational problems and low availability of the system, which currently ranges from 75 to 85% due to numerous failures. The annual number of emergency interventions exceeds 1250. The risk of system collapse is high. Annual heat losses from the DHN amount to 30–40%. With the priorities: (1) reliability and security of heat supply, (2) heat affordability, (3) heat greening, significant local resources must be engaged and large investments secured to carry out the system retrofitting and transition simultaneously.

At present, the system does not fulfil the EED criteria for EDHC because there are no renewable energy sources in use. However, the system will soon reach the status of EDHS since there are first two solar thermal systems under construction. Considering that the share of heat from CHP will be maintained at the level above the threshold of 75%, this investment will enable the system to be considered as efficient till the end of 2027. This is because EED does not provide any quantitative criterion for heat losses in DHN and the definition of EDHS is based solely on the shares of RES, WHS and CHP.

The LIFE22-CET-SET_HEAT project activities focused on heat mapping revealed almost no waste heat, and limited resources to harness renewable energy. The main identified waste heat sources are a potential municipal waste incineration plant, a wastewater treatment plant in Glina, warm ventilation air from metro tunnels, data centres, and shopping malls. The identified renewable energy sources include solar rooftop plants (total potential depicted in Figure 6), air source heat pumps (ASHPs), lake water heat pumps (WSHPs), geothermal heating plants, and biogas CHP plants using agricultural biomass and separated organic fraction of municipal solid wastes. A waste-to-energy plant incinerating what remains from waste sorting is also a strong option. In total, the newly installed heating capacity potential was estimated at around 367 MW_th, which only allows for meeting the EED’s criteria of an effective DH system until 2039. The results of heat mapping are depicted in Figure 7. The total number of identified new renewable and waste heat sources is 223, out of which 200 are the rooftops of thermal points. Further decarbonisation of the system requires effective engagement with external stakeholders and appropriate coordination of the planning activities between national and municipal authorities, utility companies, residents, and the private sector. To effectively address the DH decarbonisation challenge, local actors should consider the MultiD approach. This approach simultaneously takes into consideration many aspects, such as DH system territorial expansion, decomposition, reconfiguration, multiple distributed and intermittent heat sources, alternative decarbonised fuel-fired cogeneration, heat storage, digitisation, low-temperature heating networks, electrification and sector integration, adjustments of heat sink installations, improved energy efficiency in buildings, large-scale investments in municipal energy infrastructure, active consumers, optimisation, flexibility, resilience, integrated businesses and value stacking, democratisation, and disruption as a usual factor.

The EU Life Programme co-founded the LIFE22-CET-SET_HEAT project aims to accelerate the energy transition and decarbonisation of district heating in four targeted Eastern European countries, through triggering strategic investment programmes and a significant number of tangible projects in the field of integration of low-grade renewable energy and waste heat into second generation DHNs. The tactics to achieve real change on a large scale are based on a coordinated approach to the transition planning that prioritises collaboration, knowledge exchange, and implementation of replicable technical and non-technical solutions. Replication and standardisation are pivotal to the process; they can streamline planning, reduce costs, improve quality, facilitate communication, and ultimately accelerate investment decisions and projects’ implementation.

The central concept of the project was to develop a set of replicable model investment projects that could be adapted and implemented in various locations with minimal modifications. Based on a multi-criteria parametric assessment, eight so-called model investment projects were defined, namely:

• SET_HEAT_SEWAGE, which focuses on heat recovery from treated sewage;
• SET_HEAT_RETAIL, which focuses on heat recovery from supermarkets;
• SET_HEAT_RIVER, which focuses on a river water industrial heat pump;
• SET_HEAT_LAKE, which focuses on a lake water industrial heat pump;
• SET_HEAT_AIR, which focuses on an air source industrial heat pump;
• SET_HEAT_SOLAR, which focuses on a solar plant as a distributed heat source;
• SET_HEAT_PTES, which focuses on a remote seasonal PTES facility;
• SET_HEAT_CHP, which focuses on waste heat recovery from low-temperature cooling circuits of existing gas engine cogeneration units.

The identified types of projects were addressed with extensive pre-feasibility studies and other ready-made documentation that aimed to facilitate the development process, implementation and replication. Public versions of the pre-feasibility studies are available at the LIFE-CET-SET_HEAT project’s website [26].

The choice of investment strategy for Bucharest depends on the availability of primary energy sources, the physical feasibility of locating new technologies and integrating them into the network, and the investment and operating costs. It should also be borne in mind that, in accordance with EED requirements [1], by 2028, the share of renewable energy in total heat going into the network is at least 5% and the total share of renewable energy, waste heat or high-efficiency cogenerated heat is at least 50%. By 2035, an efficient district heating system should be using 50% renewable energy and waste heat, or 35% of either renewable energy or waste heat, assuming that at least 55% of heat will be from high-efficiency cogeneration units. These thresholds are critical for selecting investment projects. Therefore, assuming that after DHN health restoration, the annual heat consumption will be approximately 3,576,594 MWh, 5% RES share requires at least 178,830 MWh of heat from such sources. In 2035, at least 1,251,808 MWh will be required to satisfy a 35% RES share in the system.

After many discussions, it was found that technically it is possible to meet the criteria of the efficient system in 2028 and 2035. Two alternative scenarios were defined based on different energy inputs. However, taking into account the typical duration of project development, permitting implementation and procurement procedures is, in most cases, assessed at 4 to 7 years, it was concluded that regarding the starting point and the investment objectives in Bucharest, it may be difficult to meet the targets on time. Nevertheless, the pathway toward the thresholds for the efficient system set for 2035 is realistic.

Scenario 1 assumes that the key element of the transition will be a waste-to-energy project based on a waste incineration technology. The plant will be using a Refuse-Derived Fuel (RDF), which is a processed fuel made from municipal solid waste (MSW) after removing recyclables and inert materials. According to measurements the typical calorific value of RDF is 15–22 MJ/kg and 40–65% by weight are biogenic contents [27]. By defining waste and residues of biological origin as biomass, RED/RED III [28] permits energy (heat, electricity) produced from burning the biodegradable (biogenic) fraction of waste to count as “renewable energy”, provided that the waste/fuel meets the directive’s sustainability and emissions criteria. It is important to acknowledge that the heat from waste-to-energy plants is regarded as a main product, and does not fulfil the criterion of a by-product. Therefore, in this work, according to [29], it is not counted as a waste heat.

In 2019, the General Council of Bucharest Municipality adopted the city’s “Master Plan for Waste Management,” which included a provision to build a large municipal waste-to-energy incinerator. As of 2025, the authorities are still in planning/preparation mode and there is no operational decision or go-ahead for construction. Therefore, realistically, it should be assumed that the waste-to-energy plant will not be commissioned before 2030. Nevertheless, the timeframe extending up to the year 2035 is deemed to be both feasible and attainable. Similarly, projects assuming lake water and treated wastewater require long planning and implementation process. Therefore, the share of 5% RES in 2028 can be achieved by implementing solar thermal projects and air-source heat pumps (ASHPs). Eventually, the assumptions for Scenario 1 are as follows:

Implementation phase I:

I.1.

Firstly, 200 selected thermal points will be equipped with rooftop solar thermal plants. Each system will consist of 80 m² of collector aperture area, which gives around 14.1 MW of peak installed heating capacity and approximately 19,240 MWh of heat annually (It must be noted that first two rooftop solar thermal plants are being implemented and further will be deployed after confirming positive results);

I.2.

Secondly, industrial ASHPs of the total installed capacity of 29 MW will be implemented. Assuming that the ASHPs will be run for 5500 h a year, the annual heat production will be approximately 159,500 MWh.

Implementation phase II:

I.3.

Waste-to-energy plant of 235,000 tonnes/year RDF processing capacity will be deployed before 2035. Assuming 8000 h of the annual operation time, the RDF chemical energy input will be approximately 135 MW. Taking into account benchmark figures, the expected electric output will be 47.1 MW and heating output will be 67.3 MW. The annual heat generation will be approximately 538,542 MWh, assuming (with a safety margin) that 30% of RDF mass will be biomass, and approximately 161,563 MWh of heat will be regarded as renewable. In addition, it was assumed that RDF price is negative (EUR − 0.13/kg), which represents the cost of wastes utilisation.

II.1.

A lake water-source industrial heat pump of 50 MW heating capacity will be installed to harvest heat from Lake Morii.

II.2.

Geothermal heating plants of the total heating capacity of 2 MW will be installed.

II.3.

A 90 MW_th wastewater-source heat pump will be integrated with Glina sewage treatment plant (3 × 30 MW_th modular design). It is assumed that a heat pump’s output is treated as RES if its seasonal performance factor (SPF) meets a required minimum threshold.

II.4.

A municipal biogas cogeneration plant will be built of 10 MW_th heating output.

II.5.

Seasonal Pit Thermal Energy Storage (PTES) will be implemented. The total storage volume will be 1.2 million m³ (6 facilities of 200,000 m³). The storage will be integrated with the DHN without additional heat pumps, and operating temperatures of hot and cold water will be 90 °C and 60 °C.

Draft assessment of contribution from RES to heat supplied into DHN is presented in

Table 1. Assumed technical specification of the investment programme in Scenario 1.

Project	Category	Installed Heating Capacity, MWth	Potential Annual Number of Running Hours	Annual Heat Production, MWh	Contribution to the Required Annual Heat from RES in 2035
Waste-to-energy plant	Partially RES	67.3 (30% bio)	8000	538,542 (161,563 bio)	12.9%
Distributed solar plants	RES	14.1 (peak)	3000	19,240	1.5%
Lake heat pump	RES	50.0	6000	300,000	24.0%
ASHPs	RES	29	5500	159,500	12.7%
Geothermal plants	RES	2.0	8000	16,000	1.3%
Wastewater heat pump	RES	90.0	6000	540,000	43.1%
Municipal biogas CHP plant	RES	10.0	7000	70,000	5.6%
TOTAL RES		215.3	-	1,703,282	101% *

* Calculated as relative to the required RES contribution of 1,251,808 MWh.

. These assumptions were used for system modelling using the energyPRO^TM software. In addition, as a result of simulations the demand for heat storage capacity was determined.

Scenario 2 is solely based on solar thermal plants (distributed on rooftops and free parcels), water and air-source heat pumps. Phase I of this scenario is the same as in the previous case. In Phase II, however, the waste-to-energy plant is replaced by additional 14.1 MW_peak distributed solar thermal plants, and ASHPs of the total installed heating output of 14 MW_th. Providing that heat generated by a heat pump should be counted under [1] as renewable energy provided that the heat pump meets the minimum efficiency criteria [30], there will also be installed three ASHPs of total capacity of 7.2 MW to recover heat from Bucharest metro (at Unirii Station, Victoriei Station and Gara de Nord). Furthermore, two water-source heat pumps (WSHPs) of the total heating output of 1.2 MW_th will be installed to recover heat from data centres. Additionally, the total PTES storage volume was enlarged to 1.5 million m³. Initially assessed contribution of RES to heat supplied into DHN in Scenario 2 is presented in

Table 2. Assumed technical specification of the investment programme in Scenario 2.

Project	Category	Installed Heating Capacity, MWth	Potential Annual Number of Running Hours	Annual Heat Production, MWh	Contribution to the Required Annual Heat from RES in 2035
Distributed solar plants	RES	28.2 (peak)	3000	38,480	3.1%
Lake heat pump	RES	50.0	6000	300,000	24.0%
ASHPs (independent)	RES	43	5500	236,500	18.9%
ASHPs (metro stations)	RES	7.2	7000	50,400	4.0%
WSHPs (data centres)	RES	1.2	7000	8400	0.7%
Geothermal plants	RES	2.0	8000	16,000	1.3%
Wastewater heat pump	RES	90.0	6000	540,000	43.1%
Municipal biogas CHP plant	RES	10.0	7000	70,000	5.6%
TOTAL RES		231.6	-	1,259,780	101% *

* Calculated as relative to the required RES contribution of 1,251,808 MWh.

. Actual contribution of heat sources to the total production was identified after system simulations.

It should be emphasised that according to current EU legislation [1,28,29,30] all the proposed heat sources are categorised as either fully or partially RES. Furthermore, all the projects under consideration are technically feasible. In Phase 1, the focus is exclusively on projects that can be implemented prior to 2035, thereby aligning with the timeframe for the definition of EDHS, spanning from 2028 to 2034. To streamline the financial analysis, it has been assumed that all projects will be implemented in 2027.

In both scenarios, it was also assumed that three existing gas-fired cogeneration plants will be modernised by 2035. Since, within the framework of the LIFE22-CET-SET_HEAT project, significant exchange of knowledge and expertise has taken place, including study visits, it was decided that new combined-cycle gas and steam turbine plants (CCGT), as recently built in Zagreb, Croatia [31], can be replicated in Bucharest. Therefore, EL-TO Zagreb CCGT plant data were used for system modelling. Furthermore, it was assumed that by 2035, the DHN in Bucharest will be revitalised and its health restored.

Investment and operational simulations were performed for two milestone years, 2028 and 2035, corresponding to EED compliance thresholds. Each scenario was assessed under technical feasibility, financial sustainability, and environmental performance criteria. Hourly load data for 2024 were analysed alongside weather data and Romania’s day-ahead electricity market prices to evaluate the operational feasibility of industrial heat pumps. The baseline scenario is the current state of the system regarding the production assets and an assumed healthy DHN. In the cases of assumed reduced heat losses, the load duration curve analysis confirmed a peak load of 1829 MW and an average of 480 MW, corresponding to a load factor of 0.22.

As presented in [11], the transition of existing DH systems depends on the required capital expenditures (CAPEX) as well as on the market prices of fuels, electricity, and European CO₂ emission allowances (EUA). To reach acceptance, the transition plans must be financially feasible. In [11], the future market prices were anticipated until 2050. However, actual market data from subsequent years indicates that this approach may be overly optimistic. For example, EUA prices have fallen rather than risen as expected. Therefore, in this study, a simplified approach is adopted that considers constant prices as recorded in 2024. For example, Figure 8 depicts day-ahead electricity market prices in Romania in 2024 and their annual average value changes from 2021 to 2024.

The initial profitability assessment of the proposed scenarios was performed in terms of net present value (NPV), internal rate of return (IRR), and simple and discounted payback. The profitability in terms of the net present value of the project is:

$$NPV={\sum }_{y=1}^N{\displaystyle \frac{\mathsf{\Delta }C{F}_y}{{\left(1+r\right)}^y}}-CAPEX,$$

(3)

In terms of IRR, it is:

$${\sum }_{y=1}^N{\displaystyle \frac{\mathsf{\Delta }C{F}_y}{{\left(1+IRR\right)}^y}}-CAPEX=0,$$

(4)

where CAPEX stands for the total capital expenditures, ΔCF_y is the differential cash flow in the year y relative to the reference scenario, r is the discounted cash flow rate, and N is the calculation horizon (economic lifetime).

In addition, to better visualise the problem of profitability, simple (SPB) and discounted (DPB) investment expenditures payback periods were assessed using formulas:

$${\sum }_{y=1}^{SPB}\mathsf{\Delta }C{F}_y-CAPEX=0,$$

(5)

$$\sum _{y=1}^{DPB}{\displaystyle \frac{\mathsf{\Delta }C{F}_y}{{\left(1+r\right)}^y}}-CAPEX=0,$$

(6)

The financial analysis of the project was performed using a differential approach in relation to the current situation. Under the assumption that no additional revenues are generated from the sale of heat to final users, the achievement cost-effectiveness in the decarbonisation programme hinges on a reduction in the variable cost of heat production in comparison with a scenario where no action is taken. Generated operating cost savings enable the capital expenditure (CAPEX) to be recouped.

The key component of the objective function is the differential annual cash flow ΔCF_y resulting from cash flows after and before the project:

$$\mathsf{\Delta }C{F}_y={CF}_y^{\prime }-{CF}_y= -∆{\sum }_t{C}_{q,var,t}-{∆Tx}_y+{∆L}_y,$$

(7)

where CF’_y is the cash flow in the year y after the implementation of a given scenario and CF_y is the cash flow in the year y in the refence scenario, ΔC_q,var,t is the variable heat production cost reduction, ΔTx_y is the annual change in tax paid in the year y, and ΔL_y is the change in residual value of production assets in the year y.

The differential cash flow ΔCF_y is calculated against the reference “no action” scenario, which assumes no changes in heat sources but only the revitalisation of the network. It should be emphasised, that an important problem in the financial analysis of the DH system transition is the calculation of cash flows and profitability for each stakeholder, respectively. Since this is a very complex issue, in this study CF is modelled disregarding the ownership issues for all system participants, and the project control volume encompassing DHN and heat sources. Since this approach combines together production and network assets, which are in practice owned by different companies, the analysis indicates necessity for dialogue, coordination between parties and refinement.

In each hour t, the heat balance of the system must be satisfied:

$${\sum }_i{Q}_{i,t}\pm Q_{PTES,t}-Q_{loss}=Q_{DHN,in}$$

(8)

where Q_PTES is the amount of heat charged to or discharged from storage, Q_loss is the heat lost from storage, and Q_DHN is the heat supplied into DHN.

Operating windows of heat production devices are taken into consideration:

$${{\dot{Q}}_{i,min}\le \dot{Q}}_{i,t}\le {\dot{Q}}_{i,max},$$

(9)

where min and max stand for the minimum and maximum allowable heat output of the device i.

The variable part of the heat production cost takes into account changes in costs of electricity, fuel costs, and maintenance and environmental costs related to particular heat sources. In the energyPRO, the variable cost of heat is calculated based on the user-specified cost and price vectors for each hour t of a given year using the formula:

$$C_{q,var,t}={\sum }_i{\left(C_{f,i}+C_{el,i}+C_{om,i}+C_{env,i}+C_{ETS,i}-S_{el,chp,i}\right)}_t,$$

(10)

where C_f is the cost of fuel, C_el is the cost of electricity, C_om is the cost of operation and maintenance, C_env is the cost of using the environment, C_ETS is the cost of EU EUA and S_el,chp is the revenue from selling cogenerated electricity.

In each scenario, costs C_q,var,t is calculated for the given set of production units. The cost of fuel for boilers and CHPs results from heating output, energy efficiency and fuel price:

$$C_{f,t}={\displaystyle \frac{Q_t}{{\eta }_t}}{p}_{f,t},$$

(11)

where η is the energy efficiency (in the case of CHP related to heat production), p_f is the fuel price per energy unit.

Electricity costs C_el are calculated assuming the purchase at variable electricity market price and transmission and distribution costs of 50 EUR/MWh paid to the local distribution system operator (DSO). In the case of heat pumps, the cost of electricity imports results from the heating output, heat pump coefficient of performance (COP), electricity market price and electricity distribution costs:

$$C_{el,HP,t}={\displaystyle \frac{Q_t}{{COP}_t}}{(p}_{el,t}+{dc}_{el}),$$

(12)

where p_el is the electricity market price, and dc_el is the electricity distribution cost.

Equipment operation and maintenance cost for each production unit C_om is calculated using cost data from [32]. The cost of CO₂ emissions allowances C_ETS is calculated for natural gas and RDF using the heating values LHV_RDF = 21.33 MJ/kg, LHV_gas = 36.54 MJ/Nm³, and emission indices: CO₂,_ETS,RDF = 91.7 kg/GJ, CO_2,ETS,gas = 55.42 kg/GJ given in the legal regulations related to carbon balancing in the EU ETS system.

In the case of cogeneration units, the income from sales of exported electricity is assessed using the formula:

$$S_{el,CHP,t}={\sum }_{CHP}{Q_{CHP,t}\sigma }_t\left(1-\alpha \right)p_{el,t},$$

(13)

where σ is the power to heat ratio of the CHP unit, and α is the auxiliary electricity consumption of the unit.

In the case of CHP units and heat pumps working together, the electric energy is balanced within the system and costs and revenues are determined only for net imported and exported amounts.

An example of the Bucharest DH system modelled using the energyPRO v. 4.9 software is depicted in Figure 9.

One of the issues that was identified during the course of the study was the question of what approach would be most appropriate for the optimisation of production unit dispatch. The existing software tools, such as the energyPRO used in this work, have been designed to assist system planners and operators to minimise the costs under given preferences regarding production units dispatch. For example, the software is currently utilised by numerous DH companies in Denmark for optimisation of production units and storage dispatch and plant operation for a short-term period. This is performed on an hourly basis using heat production cost minimisation as the objective function and Mixed Integer Linear Programming (MILP) algorithms. The issue is that the software is not capable of implementing constraints (1) and (2) resulting from the EED directive, which apply to annual shares of RES, WHS and CHP. Therefore, in the case of considerable number of hours with relatively high electricity prices, the cost-oriented optimisation may favour operation of CHP and gas boilers. The only way to dispatch a unit when its operation is not recommended from the overall cost perspective is to define user priorities. It means, at each hour, the costs are minimised taking into consideration a user-specified priority (i.e., optimisation problem constraint) regarding operation of specific production units.

In Denmark, where the annual share of renewables in electricity production is approximately 60%, simple cost optimisation typically favours heat pumps and even electric boilers in significant number of hours. In countries like Romania, where the energy system is based on natural gas, and electricity prices are predominantly positive, cost optimisation is insufficient to prioritise electrified heat generation. Therefore, to fulfil the EED criteria for EDHS, it is essential to prioritise those production units that ensure sufficient production of heat from RES and CHP. At the current prices of fossil fuels and EU CO₂ emission allowances, this results in increased heat costs. Consequently, the issue extends beyond the high CAPEX required for new production infrastructure, to the lack of incentives for heat production from RES and WHS.

In DH systems, such as the one in Bucharest, the optimal production unit dispatch has been identified as a significant problem in real-world settings. This is due to the fact that system operators are unable to predict annual heat production in advance, and thus they optimise the system using forecasts (weather, head demand, prices) for the following 24 to 48 h. At the end of the year, it may appear that the EED requirements on the shares of particular heat sources have not been fulfilled, and consequently the system was not efficient. To prevent this, it is necessary to either develop new software tools or modify existing ones.

Key assumptions made for the study were as follows:

• The amount of heat is constant in the following years (this assumption in practice means that all heat consumption reductions will be compensated by new connections; this trend is currently observed in other LIFE22-CET-SET_HEAT project countries);
• Capital expenditures for DHN revitalisation and thermal insulation of building stock are not taken into consideration since these costs must be incurred in all scenarios.
• All financial calculations in this study were performed in constant value of money (the base year is 2024);
• Time horizon for NPV calculations is 23 years (until 2027–2050), and the constant annual cash flow is used based on the 2024 data;
• VAT is not included;
• During the investment project, operation of the DH network continues uninterrupted;
• No any form of financial support for investments is taken into account in the base financing scenario;
• As the basic financing option, the investment is financed from equity (with the share u_e = 25%) and from a bank loan (with the share u_k = 75%);
• The nominal interest rate on the bank loan was assumed to be r_k = 9.0% per annum (for corporate investments in Romania);
• The repayment period for the bank loan was set to 10 years;
• Company income tax rate (CIT) for Romania is 16% of the tax base;
• To assess the real discounted cash flow rate, the average annual inflation rate is projected to be i = 8.6% (as of October 2025 EU harmonised value);
• The nominal cost of equity was assumed as r_e = 11.24%, which accounts for a risk premium.
• The real financial discount rate r for the basic analysis was set at 10.0% based on the formula:

$$r=u_k{\displaystyle \frac{r_k-i}{1+i}}+u_e{\displaystyle \frac{r_e-i}{1+i}}$$

(14)

• The investment will not result in an increase in personnel and general administrative costs (there will be no increase in the number of jobs, and the new equipment will be operated using existing human resources);
• The straight-line method was used to determine the depreciation rate for fixed assets;
• Due to the lack of relevant data to build a model of the system services market, the calculations assumed that no such activity existed.

To assess the required capital investment costs (CAPEX), a catalogue of district heating technologies was developed [33]. Key costing curves from the catalogue are depicted in [11]. In addition, data were used from [32], from local vendors, and in some cases from engineering companies. The data from different years were indexed using the Chemical Engineering Plant Cost Index (CEPCI). It must be noted that since all of the planned energy conversion facilities will be located at the own sites of either CMTEB or partner companies, the land purchase costs were not taken into consideration. Furthermore, it was assumed that the CAPEX related to the DHN revitalisation process will be incurred anyway, and it does not influence the investments in heat sources. The summary of capital cost estimates is given in

Table 3. Anticipated capital expenditures (CAPEX) in Scenario 1.

Technology	Phase 1, Million EUR	Phase 2, Million EUR
Conversion of existing CHPs into modern CCGT plants	150–200 (1 plant)	300–400 (2 plants)
Waste-to-energy plant	0	300–400
Distributed solar plants	12.7	0
Lake heat pump	0	66.7
ASHPs	43.6	0
Geothermal plants	0	25.0
Wastewater heat pump	0	88.4
Municipal biogas CHP plant	0	40.0
Seasonal PTES (1,200,000 m3)		84.0
TOTAL estimated including 10% contingency	281.9	1214.5

and

Table 4. Anticipated capital expenditures (CAPEX) in Scenario 2.

Technology	Phase 1, EUR	Phase 2, EUR
Conversion of existing CHPs into modern CCGT plants	150–200 (1 plant)	300–400 (2 plants)
Distributed solar plants	12.7	12.7
Lake heat pump	0	66.7
ASHPs (independent)	43.6	21.2
ASHPs (metro stations)	0	11.5
WSHPs (data centres)	0	4.2
Geothermal plants	0	25.0
Wastewater heat pump	0	88.4
Municipal biogas CHP plant	0	40.0
Seasonal PTES (1,500,000 m3)	0	105
TOTAL estimated including 10% contingency	281.9	852.2

. In phase two, in addition to the proposed heat sources, the seasonal heat storage is taken into consideration. It was determined that the storage constituted a critical system component, thus facilitating the fulfilment of the EED requirements stipulated in Equation (2). Without storage, it would be necessary to significantly increase the installed capacity of particular technologies, while in summer, most of the units would be idling.

It should be noted that the financial assessment of this work is only preliminary. Electricity market prices are considered on an hourly basis, whereas gas and EUA prices are considered on a monthly basis. The annual average values used in the model were: 87.57 EUR/MWh for electricity, 37.61 EUR/MWh for natural gas, and 65.24 EUR/tonne CO₂ for EUA. In 2025, average market electricity price increased to 108.17 EUR/MWh, EUA price to 73.50 EUR, while gas prices for large non-household consumers showed relative stability with slight fluctuations, averaging around 33.00–38.00 EUR/MWh. This means, that with the electricity price increase of 23.5%, EUA price increase of 12.7%, and stable gas prices, the market favours electricity generation and CHP units. This suggests that informed investment decisions should be based on long-term market simulations, as demonstrated in [11]. As financial results are influenced by price values and price variation profiles (location of peaks and valleys), conventional sensitivity analysis was omitted in this study. The development of an appropriate methodology and the presentation of sensitivity analysis for two respective scenarios would significantly increase the volume of the text. This part was deliberately left for future work.

3. Results

The two proposed scenarios were examined in terms of system operability, energy and financial performance. Achieving appropriate ratios for the contribution of different heat sources to total heat supplied into DHN depends on their installed capacities and on their control strategies. Figure 10 depicts an example of the energyPRO simulated new production profile for scenario 1, phase 2 of the project. In the first run, the simulation was carried out to minimise variable heat costs (10). Therefore, the waste-to-energy plant was positioned as a base load unit (due to negative fuel cost), which is shown in Figure 10. However, the annual results indicated that under the specified fuel and energy prices, the thresholds (1) and (2) were not met. Accordingly, multiple simulations were conducted by modifying unit dispatch priorities to favour heat pumps over CHP units, in order to meet the required RES share in DHN heat supply. Ensuring adequate shares of heat from RES required prioritising the operation of heat pumps during periods when cogeneration units or gas boilers would otherwise have been the more cost-effective option. Key annual figures are presented in

Table 5. Bucharest DH system annual mass and energy balance results in respective scenarios.

Item	Phase 1 from 2028 to 2034	From 2035 to 2039 Scenario 1	From 2028 to 2034 Scenario 2
Total annual heat supplied to DHN, MWh	3,576,851	3,576,851	3,576,851
Heat generated by geothermal plants, MWh	0	7254	12,644
Heat generated by solar plants, MWh	17,321	15,483	35,330
Heat generated by ASHPs, MWh	162,563	124,295	245,340
Heat generated by WSHPs, MWh	0	900,957	900,718
Heat generated by biogas CHP, MWh	0	57,302	86,726
Heat generated by waste-to-energy plant, MWh	0	522,859	0
Heat generated by high-efficiency CHPs, MWh	1,927,752	1,245,934	1,542,093
Heat generated by other CHPs, MWh	68,677	9768	44,890
Share of RES in heat generated	5.0%	35.2%	35.7
Share of high-efficiency CHPs in heat generated	53.9%	34.8%	43.1
Heat generated by gas boilers, MWh	1,400,538	703,457	720.264
Heat losses from PTES, MWh	0	10,490	11,187
Total electricity generated, MWh	2,071,150	2,020,284	2,072,346
Total electricity consumed, MWh	45,830	278,030	325,864
Net electricity balance change *, MWh	+771,156	+488,091	+492,318
Electricity exported, MWh	2,037,574	1,764,897	1,907,377
Electricity imported, MWh	12,254	22,643	160,895
Total gas consumption, MWh	6,292,586	3,974,213	4,817,819
Gas consumption change *, MWh	+405,761	−1,912,612	−1,069,006
RDF consumption, MWh	0	1,045,852	0
Change in CO2 emissions *, tones/a	+80,954	−36,332	−213,280

* Calculated as onsite emissions versus the reference scenario.

. Significantly increased electricity generation can be noticed, which is the result of the conversion of cogeneration technology into modern CCGT blocks of much higher electricity-to-heat ratio. This enabled net heat production costs to be reduced compared to the reference scenario. Following a detailed analysis of the study’s findings, it has been determined that the energyPRO algorithms require refinement in order to facilitate the incorporation of constraints (1) and (2) in the context of simulations of DH systems that are currently reliant on CHP technology and gas boilers. However, it should be noted that this approach will result in significantly more computationally intensive optimisation processes.

It should be recognised that meeting the requirements of the EED depends in part on the deployment of seasonal heat storage. During the initial phase of system transition, all newly constructed production units were designed to solely address base heating loads, thereby negating the necessity for storage. However, from 2035 onward, the storage facilitates the transfer of heat from the summer period to the heating season. In the optimisation model, the operation of PTES is oriented towards system balancing and supporting the operation of production units on an hourly basis. The state of charge is depicted in Figure 11, which reveals that several charging and discharging periods occur, and that the functionality of the storage extends beyond the seasonal. It is noteworthy that in transition scenario 1, the storage’s functionality within the system is more akin to seasonal storage, while in scenario 2, its role is more oriented towards system balancing.

In relation to the reference scenario, electricity generation increased. This is the result of the conversion of old CHP plants into new CCGT blocs. In the reference scenario, due to unfavourable performance indicators, for many hours, gas boilers were prioritised over CHPs. After retrofitting and a technology change, CCGTs became key contributors to heat generation, and the operation of gas boilers was largely eliminated.

In phase 1, distributed solar plants contribute to the annual heat supplied into DHN with 0.5%, ASHPs with 4.5% and high-efficiency CHP with 53.9%, which fulfils the EED requirements for EDHS. However, in phase 2 of both scenarios, although more than 35% of heat supplied into DHN is from RES, the total share of high-efficiency CHP, RES and WHS is below the required 80%. Moreover, share of boilers was still considerable. This is an important issue of transition planning, that cannot be overlooked.

In this work, due to the inability of the optimisation software to include constraints (1) and (2) directly in the optimisation task, the gas boiler operation priority was set to low to increase the share of high-efficiency CHP. This means that the boilers are dispatched if other units cannot balance the system. The results revealed that such a constraint practically shifted heat production to CHP, even if electricity prices were not favourable. As a result, electricity production, gas consumption and total CO₂ emissions increased considerably. The mass and energy balance for the changed system operation, and considering different unit dispatch priorities is given in

Table 6. Bucharest DH system annual mass and energy balance results with eliminated gas boilers.

Item	From 2035 to 2039 Scenario 1	From 2028 to 2034 Scenario 2
Total annual heat supplied to DHN, MWh	3,576,851	3,576,851
Heat generated by geothermal plants, MWh	7258	12,648
Heat generated by solar plants, MWh	15,483	35,330
Heat generated by ASHPs, MWh	125,737	247,399
Heat generated by WSHPs, MWh	363,022	901,273
Heat generated by biogas CHP, MWh	57,295	86,726
Heat generated by waste-to-energy plant, MWh	522,917	0
Heat generated by high-efficiency CHPs, MWh	1,725,481	2,047,630
Heat generated by other CHPs, MWh	232,535	257,370
Share of RES in heat generated	35.2%	35.8%
Share of high-efficiency CHPs in heat generated	48.1%	57.1%
Heat generated by gas boilers, MWh	0	0
Heat losses from PTES, MWh	10,793	11,560
Total electricity generated, MWh	2,697,798	2,780,385
Total electricity consumed, MWh	278,539	326,748
Net electricity balance change *, MWh	+1,165,096	+1,199,473
Electricity exported, MWh	2,441,902	2,565,667
Electricity imported, MWh	22,642	112,030
Total gas consumption, MWh	4,820,123	5,692,043
Gas consumption change *, MWh	−1,066,702	−194,783
RDF consumption, MWh	1,045,951	0
Change in CO2 emissions *, tones/a	132,469	−38,862

* Calculated as onsite emissions versus the reference scenario.

.

The final values of RES and CHP shares in the total amount of heat supplied into the DHN are shown in

Table 7. RES and CHP shares in the total amount of heat supplied into the DHN.

Heat Source	From 2028 to 2034	From 2035 to 2039 Scenario 1	From 2028 to 2034 Scenario 2
Geothermal plants	0.0%	0.2%	0.4%
Solar plants	0.5%	0.4%	0.9%
ASHPs	4.5%	3.5%	6.9%
WSHPs	0.0%	25.0%	25.0%
Biogas CHP	0.0%	1.6%	2.4%
Waste-to-energy plant *	0.0%	4.4%	0.0%
High-efficiency CHP	53.9%	48.2%	57.2%
Total share of RES in heat supplied into DHN	5.1%	35.1%	35.6%
Total contribution of eligible heat sources	59.0%	83.3%	92.8%

* Calculated under assumption that only biogenic fraction of RDF is 30% by mass.

. It should be noted that these values represent net values after considering heat losses from storage, and are in accordance with the criteria for EDHS as defined in the EED. The calculations demonstrate that the proposed portfolio of investment projects and technologies is capable of ensuring compliance with legal regulations. It is imperative to emphasise that, post-2035, the value of shares in specific categories of heat sources will exhibit minimal safety margins, thereby constricting the scope for system optimisation and flexibility. Consequently, the operation of particular units must be meticulously monitored to ensure compliance with regulatory thresholds.

Table 8. Key financial operating results.

Investment Scenario	Total Operating Expenditures, EUR	Revenues—Sales of Electricity, EUR	Net Differential Operating Cash Flow, EUR
Reference scenario	325,341,051	175,366,522	-
Scenarios 1 and 2, Phase 1	366,218,187	269,815,460	−53,571,803
Scenario 1, Phase 2	259,283,456	243,731,733	−134,422,807
Scenario 1, Phase 2, EED satisfied	318,458,194	291,715,609	−123,231,945
Scenario 2, Phase 2	339,458,678	272,275,872	−82,791,724
Scenario 2, Phase 2, EED satisfied	397,914,280	318,607,991	−70,668,240

summarises the key financial effects of the decarbonisation programme for each of the considered scenarios. It is noteworthy, that during the second phase of the transition programme implementation, the need to adhere to the EED resulted in a deterioration of the net differential operating cash flow. Expected values of profitability indices are given in

Table 9. Estimated values of profitability indices.

Index	Scenario 1	Scenario 2
NPV, million EUR	−195.1	−201.9
IRR	0.056	0.036
SPB, years	16.2	18.3
DPB, years	Not found within the given time horizon	Not found within the given time horizon

. Figure 12 depicts the change of the simple and discounted values of the project over th eperiod from 2027 to 2050.

4. Discussion

The analysis confirms that achieving compliance with the Energy Efficiency Directive (EED) targets for 2028 and 2035 in the Bucharest district heating (DH) system is technically feasible. However, this feasibility is strongly constrained by system characteristics and local boundary conditions. The system’s second-generation design, high supply temperatures, extensive heat losses, and limited availability of local renewable and waste heat resources within the city boundaries significantly restrict the range of effective decarbonisation options.

A central finding is the strong sensitivity of EED compliance to operational strategy. When heat production is dispatched solely on the basis of minimum variable costs, the proposed system configurations fail to meet the required RES shares. Conversely, enforcing compliance by prioritising the operation of heat pumps and CHP units leads to higher operating costs, increased electricity consumption, and higher CO₂ emissions. This demonstrates a fundamental trade-off between economic optimisation and regulatory compliance in large legacy DH systems.

Seasonal thermal energy storage plays an important role after 2035 by enabling the temporal shift of renewable heat from summer to the heating season. However, the required storage volumes are substantial, and realistic deployment is limited by land availability, investment costs, and integration complexity in a dense metropolitan context. As a result, dispatchable low-carbon heat sources remain indispensable for maintaining system reliability during peak demand.

The comparison of scenarios indicates that the waste-to-energy-based pathway offers greater operational stability and lower variable costs. However, it depends on long lead times, permitting processes, and waste policy decisions that lie beyond the control of the DH operator. The alternative pathway relying exclusively on heat pumps and solar thermal increases exposure to electricity price volatility and requires larger storage capacities.

From a financial standpoint, both scenarios result in negative net present values under conservative assumptions in the absence of external support. This substantiates the conclusion that the deep decarbonisation of large, capital-intensive DH systems cannot be solely financed through operational savings. Given that the transition programme is not financially viable in the long term, with strongly negative NPVs for both proposed scenarios, the results underscore the necessity for either significant subsidies or an increase in heat prices for final consumers to boost incomes. Other potential sources of income, such as auxiliary services for the electricity grid and heating services, are also proposed. However, these must be developed in Romania. At present, the financing of the transition programme and its profitability are critical issues.

From a policy perspective, the study highlights that Bucharest’s transition requires a coordinated governance model integrating all relevant stakeholders, including municipal authorities. Institutional fragmentation remains a critical bottleneck and continues to delay investment implementation. Decarbonising Bucharest’s DH system requires appropriate policy instruments, unprecedented investment levels, and effective coordination. The study highlights that without EU-level financial support and tariff reforms, achieving cost-effective decarbonisation remains unlikely.

5. Conclusions

The EED poses a serious challenge to large DH systems in terms of decarbonisation. This study presents a strategic, system-level assessment of decarbonisation pathways for the Bucharest district heating system and offers insights relevant to other large, fossil-fuel-dependent second-generation DH systems. The results show that compliance with the EED criteria for 2028 and 2035 is achievable in principle, but only under restrictive technical and economic conditions.

The main conclusions are as follows:

• Compliance with EED targets is not equivalent to system cost efficiency and may require operational compromises that increase costs or emissions.
• Dispatchable low-carbon heat sources are essential to ensure security of supply and limit overreliance on electricity-driven technologies.
• Seasonal thermal energy storage is a key enabler, but its feasible scale is limited in dense urban environments.
• Modernised CHP plants remain critical transitional assets, supporting flexibility and mitigating electricity market risks.
• Financial viability cannot be achieved without external support, underscoring the necessity of EU funding, regulatory incentives, and risk-sharing mechanisms.

The analysis shows that compliance with the EED requirements for 2035 can be achieved with relative ease through a targeted portfolio of renewable energy, waste heat, and high-efficiency cogeneration technologies. However, the results also indicate that further progress toward deep decarbonisation becomes significantly more challenging beyond this milestone. As the most accessible and dispatchable low-carbon resources are exhausted, achieving additional emission reductions becomes more difficult. Further progress increasingly depends on complex system reconfiguration, large-scale energy storage, electrification, and extensive coordination across sectors and stakeholders. This confirms that incremental, technology-by-technology approaches are insufficient and that multidimensional, holistic strategic planning is indispensable. Such planning must simultaneously address network transformation, heat demand reduction, market integration, governance structures, and social acceptance. A common feature emerging from both this study and comparable transition strategies is the growing emphasis on energy harvesting, systematic recovery of low-grade renewable and waste heat dispersed across the urban environment, which increasingly defines the backbone of long-term district heating decarbonisation pathways.

In addition to the modernisation of existing production facilities and networks, and the integration of multiple distributed heat sources, it is recommended that future activities concentrate on transitioning to fourth-generation (4GDH) standards and gradually reducing the temperature in the DHN. Network performance studies indicate that Bucharest has considerable potential for this transition, given the oversized pipeline diameters and the temperature difference of 30–40 K between the primary and secondary networks. This has, however, not been addressed in this study and left for future work.

The socio-economic dimension is equally critical. Residential heat tariffs are heavily subsidised, creating a fiscal burden on the municipality but protecting low-income consumers. This policy complicates investment cost recovery and discourages energy efficiency. At the same time, heat affordability remains politically sensitive, making social dialogue essential for a just transition.

Overall, the decarbonisation of large DH systems should be understood as a multidimensional socio-technical transition rather than a purely technological upgrade. Effective transition strategies must integrate infrastructure renewal, system flexibility, spatial planning, governance coordination, and long-term policy alignment.

Future work will focus on uncertainty and sensitivity analyses of energy and CO₂ price trajectories, detailed modelling of network temperature reduction and building-side adaptations, assessment of business models for waste heat integration, and broader environmental and social impact evaluations. Further research will also examine the replicability of the proposed methodological framework across other metropolitan DH systems in Central and Eastern Europe.