Energy, Economic, and Environmental Assessment of Wind Turbine Blade Thermal Recycling Coupled with Organic Rankine Cycle Heat Recovery and Power Generation

Abstract

Wind turbine blade (WTB) end-of-life waste is projected to increase significantly, yet no sustainable recycling solution with a clear energy, economic, and environmental (3E) assessment exists. This paper presents a validated 3E model of a WTB thermal recycling pilot (1 t/day) to benchmark recycled glass fibre (rGF) against virgin glass fibre (vGF) and identifies the throughput at which rGF becomes competitive. This subsequently leads to a projection of 3E performance at 5000 t/y plant capacity, at which rGF achieves approximately 46% lower specific primary thermal energy, 92% of the CO₂ emissions of vGF, and a selling price of 80% of vGF for a financial break-even. Building on this baseline, a novel combined material, heat, and power system is proposed and simulated, integrating the WTB recycling pilot with a 20 kWₑₗ/130 kWₜₕ organic Rankine cycle to serve residential buildings. Results show that coupling the pilot with 3000 m² of apartments yields a near net-zero CO₂ and energy-cost residential complex, with overall CO₂ emissions falling below those of standalone residential buildings combined with vGF production when more than 25 apartments are integrated.

1. Introduction

Wind energy is one of the most promising and rapidly developing renewable energy sources, capable of replacing a significant share of fossil fuels in electricity production [1], and is recognised for its potential to reduce greenhouse gas emissions, thereby contributing to a more sustainable energy future [2]. The global wind energy capacity has seen exponential growth, with installations reaching over 743 GW by the end of 2020 [3]. This growth is attributed to both onshore and offshore wind farms, which benefit from the higher wind speeds and more consistent wind patterns found over open water [2]. The efficiency of wind turbines has also improved, with modern turbines capable of converting up to 45% of the wind’s kinetic energy into electricity.

Despite its benefits, wind energy faces challenges such as intermittency, which requires hybridisation and integration with other energy sources or storage systems to ensure a stable power supply [4,5]. Additionally, the environmental impact of wind farms, including effects on wildlife and noise pollution, necessitates careful site selection and technological innovations to mitigate these issues [6,7].

A more recent concern is the absence of a sustainable strategy for disposing of end-of-life wind turbine blades (WTB), as some of the earliest installed wind farms are approaching their design lifetime of approximately 20–25 years, depending on the maintenance and environmental conditions [8]. The global volume of the WTB waste is projected to surge rapidly, as estimated by Liu and Barlow [9] from approximately 500,000 tonnes between 2029 and 2033 to 43 million tonnes by 2050, assuming a moderate growth rate. WTB waste estimates reported in the literature are not always consistent and should be updated in the coming years, as they are subject to variation due to blade life extension and changes in WTB owners’ policies.

Several measures can be implemented to reduce waste at the source, such as blade life extension and fault diagnosis to schedule preventive maintenance. While these measures can help to some extent in reducing peak waste generation, a large volume of this waste still requires sustainable management. Landfilling is currently the least costly solution for end-of-life WTB, making it the most common option, followed by incineration [10]. However, landfilling is becoming increasingly challenging owing to limited land availability and growing regulatory constraints related to sustainability. Furthermore, landfilling results in the loss of all valuable and recoverable material components. The manufacture of WTBs involves significant energy consumption and CO₂ emissions, which must be accounted for in the life cycle assessment of wind farms to accurately quantify the environmental impacts of wind energy systems. Hence, sustainable recycling solutions are required to manage the large volume of WTB waste and recover valuable materials such as glass fibre (GF).

Recycling is a challenging yet necessary process in many sectors, delivering significant economic and environmental benefits within the framework of the sustainable and circular economy. Recycling is one of the few domains in which financial interests align with the reduction in ecological impacts. Among feasible and economically viable recyclable materials, WTB waste is particularly significant owing to its rapidly increasing volume and the current absence of a sustainable solution capable of handling it at such scales. While 80 to 94% of a wind turbine can be recycled, including metallic components [10], recycling the blades remains a significant challenge. The blades are difficult to recycle, as they are composed of more than 90% glass-fibre reinforced polymer (GRP) [11] consisting of 60–70% GF, approximately 30% thermosetting epoxy resin, and low concentrations of other materials such as balsa wood, PVC, paint, putty, and adhesives. The robust structural integrity of the blade and the irreversible cross-linking properties of the thermoset matrix make WTB recycling challenging [12]. Therefore, a viable, sustainable recycling solution must be capable of handling large throughputs at reasonably low overall costs and carbon footprint, whereas existing solutions remain limited in capacity, irrespective of sustainability considerations. A suitable recycling solution would contribute to a more circular economy for GF and the WTB industry.

A sustainable recycling solution produces recycled GF (rGF) as the main product and potentially some carbon fibres, metals, and liquid hydrocarbons. Alternatives to WTB recycling include blade reuse [13] and repurposing [14,15]; however, these options offer limited capacity. Recycling methods have been developed based on mechanical [16], thermal [17], and chemical processes [18], as well as combined approaches [19] to potentially address the large volume of the glass and carbon fibre composite waste [20].

Mechanical recycling breaks down the blades into flakes, pellets, and granulates through shredding, crushing, and milling. The resulting products are used as reinforcement in composites such as BMC, cement [21,22], and asphalt [23]. It is the least costly, scalable recycling method and serves as the basis for other methods to some extent. However, it does not produce rGF, and the cost of the resulting GRP fillers is difficult to justify against the existing mineral fillers.

The main thermal processes are pyrolysis [24], gasification, fluidised-bed combustion [17], and cement co-processing [25]. In the pyrolysis process, composites are heated in the absence of oxygen, producing rGF/rCF, pyrolysis gas, and liquid hydrocarbons; it is the most widely used method to date, particularly for recycling carbon fibre products [26,27]. In the gasification process, materials are heated in a controlled, oxygen-lean environment; this process has not yet matured into an established technology for recycling fibre-reinforced polymers [28]. The main products are rGF/rCF, and the process gas consists mainly of H₂ and CO. In both pyrolysis and gasification processes, the recycled fibres are contaminated with residual char on the surface, and an afterburner is required to oxidise the surface char [29]. In the gasification process, materials are heated in a controlled, oxygen-lean environment. The cost of these process struggles to reconcile with the economic requirements for rGF, given the low cost of virgin GF (vGF) [30].

The main product of the fluidised bed process is the rGF/rCF, with the heat as a by-product of the combustion of the polymer content of the composites. Fluidised-bed reactors provide a uniform temperature distribution in the reactor and are tolerant of mixed materials and contamination [31]. This process is less energy-intensive than pyrolysis, owing to the full combustion of the polymers; however, the tensile strength of the produced rGF/rCF is more adversely affected than in pyrolysis at the same temperature due to the combined effects of oxidation and attrition at high temperatures. Cement co-processing is the most cost-effective and scalable method [10], reducing the CO₂ emissions and water consumption by 27% and 13%, respectively, as reported by Quantis U.S [32]. However, the fibres are downcycled as they are used as fillers in the cement. Moreover, cement kilns require a large and consistently available supply of downsized GRP waste with adequate calorific value to incorporate these materials at scale, which is not feasible in many regions, such as the UK [33].

Solvolysis [18,30] and pressolysis [34] are examples of chemical processes that use reactive solvents to decompose the polymer and reclaim the fibres. The principal advantages of the thermal processes over chemical processes are their scalability, lower cost, and shorter processing times. Recycling cost is particularly critical for GF, given the low market price of vGF. Furthermore, GRP materials, including WTBs, are composed of multiple constituent materials rather than pure GF and thermoset polymer alone, which poses a challenge for chemical processes. However, the tensile strength of rGF is reduced by thermal methods, particularly in fluidised-bed reactors, whereas the degradation can be more limited in the chemical methods. Reviews of various aspects of the 3E performance of common composite recycling methods are presented in [19,20,35,36,37], while the present work conducts a 3E assessment based on a case study and a purpose-built demonstration plant developed within the PRoGrESS project [33].

In addition, research and development are ongoing on more sustainable manufacturing methods for future WTB through the development of recyclable polymers, enabling more straightforward recovery of both the polymer and GF [38,39]. Moreover, WTB recycling methods are generally applicable to a wide range of fibre-reinforced products, affording greater supply chain flexibility.

Accordingly, a demonstration plant for the WTB thermal recycling process using fluidised-bed technology was developed within the PRoGrESS project [33], representing the first pilot of its kind. The recycling pilot (1 t/day) modelled in this work represents the first stage of upscaling towards industrial-scale operation. The pilot is modelled based on a calibrated model of the demonstration reactor developed using the experimental data obtained from it. Fluidised-bed reactors used for composite recycling in the literature have operated at temperatures of 450–550 °C using silica sand as the fluidising medium [29]. The rCF and rGF produced in these reactors exhibited approximately 20% and 50% reductions in tensile strength, respectively, while the original stiffness was retained [29,33]. The typical mean fibre length (by weight) in fluidised-bed reactors ranges from 6 to 10 mm [29].

Fibre length is one of the key parameters governing the performance of fluidised-bed composite recycling, affecting downsizing costs, reactor throughput, feedstock residence time in the reactor, fibre strength, and the mechanical properties of GRP composites manufactured with rGF. Pickering et al. [40] demonstrated that the feeding rate of carbon fibre products into a fluidised-bed reactor increased exponentially with decreasing fibre length, at constant temperature and air velocity in the reactor. Consequently, reactor throughput increases in a similar manner for smaller feedstocks. An increase in reactor throughput per unit of reactor area reduces the total energy consumption per unit mass of recycled fibre produced [41]. Consequently, the carbon footprint of the recycled fibres decreases with increasing reactor throughput.

Fibre length is also critical to the re-processability and performance of secondary composites incorporating rGF. Short fibres tend to exhibit less entanglement, facilitating easier re-processing and higher tensile strength owing to the size effect. They are therefore well-suited to composite applications involving BMC and the injection moulding process. Furthermore, certain fibre realignment processes can also benefit from short fibres owing to their easier dispersion and alignment. Conversely, long fibres can potentially yield superior composite performance owing to more effective load transfer, and certain composite applications involving textile processes require a certain minimum fibre length. Therefore, the feedstock size, and by extension the length of recycled fibres, represents a critical trade-off between various performance indicators spanning the entire waste processing cycle (e.g., downsizing, transportation, and recycling) through to the reuse stage in terms of re-processability and composite performance [42].

WTB waste is a promising feedstock for fluidised bed technology owing to its low resin and high GF contents [43]. The energy, economic, and environmental (3E) performance indices of rGF production and WTB recycling are assessed in this work and compared with those of vGF production and WTB landfilling. Favourable 3E indices for WTB recycling and rGF production are essential for developing and promoting sustainable WTB waste management and establishing a circular economy for GF.

Waste heat recovery holds significant potential for improving the efficiency of industrial plants, thereby reducing their fossil fuel consumption and associated carbon emissions [44]. Organic Rankine cycle (ORC) systems have attracted particular attention in recent years due to their reliability, flexibility, safety, and applicability for low-temperature power generation [45,46]. ORC systems are among the most promising technologies for converting low-grade heat into electricity [47]. ORC systems have been investigated in the literature and have demonstrated high reliability through commercial trials. For example, small-scale ORC systems offer an extended operational lifetime of up to 30 years, owing to the use of dry working fluids and the absence of liquid-phase transitions, compared with 15–20 years for steam turbines [48]. This also results in straightforward and cost-effective maintenance with no major overhaul required, as stated by the world’s largest ORC manufacturer, Turboden [49]. Turboden also reported an average availability of more than 98% across all installed units [50], demonstrating their high reliability and operability. Furthermore, the high reliability and operability of the ORC systems were corroborated by Turboden’s authorised ORC distributor, INMIS Energy, which schedules planned maintenance only once or twice per year [51].

In addition, ORC systems have shown promise for combined heat and power (CHP) applications owing to their simplicity and flexibility in adapting to existing conventional heating systems in residential dwellings [52]. The building sector is considered the target application, as it accounts for approximately 40% of the EU energy consumption and 36% of the greenhouse gas emissions [53]. Households in the EU accounted for 27% of final energy consumption in 2020 [54], of which 62.8% was used for space heating, 15.1% for water heating, and 0.4% for space cooling.

Combined production is an effective approach to promoting sustainable energy solutions and valorising processes and power systems through the integration of processes, energy vectors, and end users [55]. The sustainability gains achieved through cogeneration are realised by improving energy efficiency, resulting in consistent energy conservation and avoidance of the associated CO₂ emissions [56]. While polygeneration systems [57] have been widely studied for the simultaneous production of multiple energy vectors [58], no prior work has extended this concept to include a recovered solid material as a co-product alongside heat and power. The present work proposes such a framework, termed here as combined material, heat, and power (CMHP), in the context of WTB thermal recycling coupled with ORC-based residential cogeneration. ORC-CHP systems coupled with similar processes and applications have been investigated in the literature and have demonstrated significant improvements in system 3E performance. Behzadi et al. [59] simulated a waste-to-energy plant in Tehran integrated with an ORC unit. They concluded that the ORC-CHP plant significantly improved exergetic and economic performance, with the ORC recovering otherwise-wasted flue gas heat to generate additional electricity while reducing the overall system cost rate, demonstrating that thermal waste processing and ORC cogeneration are thermodynamically and economically synergistic. Braimakis et al. [60] conducted a techno-economic analysis of a small-scale biomass ORC-CHP system serving residential buildings of varying community sizes across different European climate zones. They found that the economic viability of the system is highly sensitive to biomass and heat selling prices, community size and heat demand profile, with larger connected building stocks significantly improving the payback period.

A novel CMHP generation system is proposed and modelled in this work. The CMHP system is distinguished by its use of WTB waste as the input material, producing rGF as the primary output, and heat and power as valuable by-products through the integrated ORC-CHP system. The modelled integrated system consists of a WTB thermal recycling pilot (1 t/day) and an ORC unit coupled with the residential buildings to supply their thermal and electrical demands in Aberdeen, Scotland. The thermochemical process model of the WTB recycling pilot is based on a calibrated 1-D model of a demonstration plant. The pilot results are used to estimate the 3E performances of the scaled WTB recycling facilities, enabling a fair comparison with established vGF manufacturers. The novelties of this work relative to the state of the art are as follows:

• A novel CMHP system is proposed, comprising an innovative WTB thermal recycling pilot based on the fluidised-bed technology coupled with an ORC-CHP system.
• A detailed model of a WTB recycling pilot (1 t/day) is developed based on experimental data from a demonstration plant and used as the baseline for scaling the process.
• The primary energy consumption, CO₂ emissions, and production costs of rGF from scaled WTB recycling units across a wide range of throughputs are presented and compared with those of vGF.
• The impact of GRP recycling and the utilisation of rGF produced by the WTB recycling pilot in BMC manufacturing is investigated and compared with the use of vGF in terms of manufacturing costs and total CO₂ emissions.
• The annual 3E performance of the residential buildings is investigated, and the impact of integrating with the proposed CMHP system is assessed.
• The paper is structured as follows: the models of the individual units of the CMHP system are presented first. Subsequently, the primary thermal energy consumption, CO₂ emissions, and the economics of rGF produced from WTB recycling facilities are presented at scaled capacities, considering both the Scottish and average UK energy grids, and compared with those of vGF. Finally, the results for the integrated residential buildings and the WTB recycling pilot coupled through an ORC-CHP system are presented and compared with those of stand-alone units.

2. Modelling of the CMHP System

The schematic of the integrated CMHP system is shown in Figure 1. It consists of five main units: (I) the WTB thermal recycling pilot (fluidised-bed technology), (II) the ORC system, (III) the water heating/cooling unit, IV) the electricity grid, and V) the residential buildings.

The WTB pilot is a thermal recycling process that accepts the shredded WTB as the feedstock and produces short rGF as the main product, with the heat generated from the combustion of the epoxy resin serving as a byproduct to be utilised by the downstream ORC system. The WTB material is shredded and downsized into 1–10 mm flakes suitable for the recycling process, the flake size being directly related to the length of the recovered rGF produced. The process requires thermal and electric energy supplied by the electrical grid to process the WTB material.

The waste heat from the WTB recycling pilot is recovered by a small-scale ORC unit to generate electricity and heat, as shown in Figure 1. The generated electricity supplies the users’ electric demand, and the excess electricity is exported to the electricity grid, while the grid is available to users when on-site production is insufficient. The thermal output at the condenser of the ORC system is directed to the water tank and passes through an internal spiral tube heat exchanger (HE) unless the tank temperature has reached its pre-set maximum setpoint, at which point the cooling water is redirected to the dry cooler (DC) via the bypass line. The tube HE is modelled using a fixed overall heat transfer coefficient (OHTC), and the model requires the tank and coil geometry as inputs. In contrast, the hot water (HW) and the domestic hot water supply for space heating (SH) are modelled using an energy balance applied to the water mass in the tank at each time step, without explicitly modelling the HE and the heat transfer within it. If the tank temperature is below the pre-set threshold, the tank is bypassed, the required thermal energy is supplied by the boiler, and the tank is charged using hot water from the ORC condenser. The tank temperature is therefore calculated at each time step using an energy balance accounting for the thermal energy output of HW and SH, thermal losses, and the thermal energy input from the ORC condenser. The models of the main units and their components are presented in the following subsections in detail.

2.1. WTB Fluidised-Bed Recycling Pilot

The WTB flakes decompose in the fluidised-bed reactor at approximately 500 °C. The epoxy resin combusts in the reactor, and the liberated rGF is elutriated out of the reactor and separated from the combustion gases using cyclone separators. The products are short and fluffy rGF that can be reused in the manufacture of less-demanding applications such as BMC and thermoplastic injection moulding. This represents a feasible strategy, given that the recycling process degrades the mechanical properties of rGF and yields short fibres, making a return to the original WTB application—which demands high-quality, continuous GF—impractical. The recycling plant can also process other types of GRP waste manufactured with thermoset resins.

The WTB thermal recycling pilot has a capacity of 1 t/day and employs a fluidised-bed reactor. Its configuration corresponds to the first intermediate step on the prospective commercialisation route of the pilot plant within the PRoGrESS project [33]. The pilot is more energy-intensive with higher operational costs and carbon emissions than the large-scale commercial plants, as has also been observed in [61]. The system is modelled on a calibrated model of a demonstration recycling unit, developed from experimental data. The model of the WTB recycling demonstration plant and the pilot (1 t/day) are implemented in Aspen Plus^® (version 14) [62]. The model of the demonstration reactor is first set up to calibrate the kinetic parameters governing epoxy decomposition inside the reactor. Accordingly, the thermochemical model of the reactor is a 1-D model with the process gas composition and the reactor energy balance varying along the reactor height. The kinetic parameters are calibrated so that the predicted gas composition leaving the reactor and the corresponding species match the experimental data recorded from the demonstration reactor at the same operating conditions. The mass flow rate of natural gas is calculated by the solver to satisfy the energy and mass balance of the reactor and maintain its temperature at approximately 500 °C. The developed model of the recycling pilot is used as the baseline for the 3E assessment of the scaled recycling facilities.

The main energy, economic, and environmental parameters of the pilot plant’s model used in the assessment of the integrated system are presented in Table 1. The plant’s CO₂ output is based on the direct emissions arising from the decomposition of WTB material in the process and natural gas (NG) consumption on site, and the indirect emissions associated with electricity consumption at the site, calculated using the Scottish and average UK electricity grid carbon intensities in 2021 [63]. The energy rates for buying and selling electricity are representative capped values; they may vary by time, consumption, application, tariff plan, and energy provider. Accordingly, typical values in the UK market are considered here without any incentives.

The WTB recycling gate fee is treated as a variable and is assumed to fall within a range relative to the landfilling fee. While predicting the gate fee for WTB recycling remains difficult at this stage, it is reasonable to assume that it would exceed the landfilling fee, given that it offers a sustainable disposal route for WTB waste owners. The rGF selling price is estimated for different recycling capacities at which the break-even point is reached. The rGF selling price is a function of multiple variables and strongly depends on market trials and the degree of rGF adoption in bulk product manufacturing, which has yet to be established.

The recycling operating expenditure (OpEx) encompasses primary energy consumption (NG and electricity), maintenance, depreciation, overheads, labour, shredding costs, and total sales, general, and administrative (SG&A) expenses. The SG&A term includes research and development (R&D), general and administrative costs, sales and marketing, and distribution costs. These parameters are calculated according to the values recommended by Fonte and Xydis [64] for a recycling plant with a 6000 t/y capacity, which are reported in Table 1. The shredding cost is estimated based on trials within the PRoGrESS project conducted to downsize 5–10 m WTB sections to 1–10 mm granulates [33]. The capital expenditure (CapEx) is estimated based on the actual construction costs of the WTB recycling pilot within the PRoGrESS project [33] and the industrial rule of thumb for costing the construction of process plants. The resulting 3E performance indicators of the recycling pilot (1 t/day) are reported in Table 1.

Table 1. The input variables to the economic and environmental model of the WTB recycling pilot (1 t/day), and the obtained 3E results from the model.

A—Energy Prices
Parameter	Value	Unit	Notes
Electricity purchase price	34.0	p/kWh,el	[63]
Electricity export price	7.5	p/kWhel,ex	Selling to grid [65]
NG purchase price	10.3	p/kWh,th	[63]
Standing charge—electricity	53.0	p/day,el	[63]
Standing charge—NG	29.0	p/day,th	[63]
Grid thermal equivalent	2.5	kWh,th/kWh,el	Assuming grid efficiency 40% [66]
B—Emissions Factors
Parameter	Value	Unit	Notes
CO2 intensity—Scottish grid	26.9	g·CO2/kWh,el	[63]
CO2 intensity—UK grid	153.6	g·CO2/kWh,el	[63]
CO2 intensity—vGF	1.9	kg·CO2/kg·vGF	Average of 10 reliable data points in [67]
CO2 intensity—polyester	7.6	kg·CO2/kg·polyester	[68]
Embodied thermal footprint—vGF	22.5	MJ,th/kg·vGF	Average of values reported in [69]
C—Material & Waste Costs
Parameter	Value	Unit	Notes
Selling price—vGF	1.7	£/kg·vGF	[64]
Selling price—rGF	1.36	£/kg·vGF	Assuming 80% of the vGF price
Selling price—polyester	3.0	£/kg·polyester	[68]
WTB landfilling gate fee	150.0	£/t	[33]
WTB shredding cost	580.0	£/t	From 5–10 m chunks to 1–10 mm flakes [33]
D—Capital & Operating Costs
Parameter	Value	Unit	Notes
Total CapEx	1071.4	k£	Incl. 20% contingency to consider CHP costs
Labour cost	131.4	k£/y	1 person/shift, 3 shifts, labour fee of 15.0 £/h
Maintenance	64.3	k£/y	6% of CapEx [64]
Depreciation	107.1	k£/y	10% of CapEx [64]
Overheads	113.4	k£/y	25% of OpEx incl. energy, labour, and maintenance [64]
Total SG&A	236.0	k£/y	35% of OpEx excl. shredding costs [64]
E—Plant Operating Parameters
Parameter	Value	Unit	Notes
WTB throughput capacity	365.0	t/y	The pilot
Annual operational hours	8760	h/y
GF content of WTB feedstock	67.7	wt%	[33]
E—Results of the Produced Recycled Glass Fibres
Parameter	Value	Unit	Notes
rGF production capacity	247.1	[t/y]	Considering 67.7% of WTB GF, and the rest epoxy
Embodied thermal footprint—rGF	15.03	MJth/kg rGF
Embodied electric footprint—rGF	6.49	MJel/kg rGF
Embodied equivalent thermal footprint—rGF	31.25	MJth,eq/kg rGF	By converting the electric footprint to a thermal footprint
CO2 intensity—rGF	2.24	kg CO2/kg rGF	For the plant’s location in Scotland
OpEx	1121.9	[k£/y]	Total OpEx incl. energy, labour, and maintenance, depreciation, overheads, shredding costs, and SG&A
Gross cost—rGF	4.54	£/kg rGF	Total OpEx, excl. revenues from selling rGF and gate fee
Net cost—rGF	2.96	£/kg rGF4	Total OpEx, incl. revenues

Table 1. The input variables to the economic and environmental model of the WTB recycling pilot (1 t/day), and the obtained 3E results from the model.

A—Energy Prices
Parameter	Value	Unit	Notes
Electricity purchase price	34.0	p/kWh,el	[63]
Electricity export price	7.5	p/kWhel,ex	Selling to grid [65]
NG purchase price	10.3	p/kWh,th	[63]
Standing charge—electricity	53.0	p/day,el	[63]
Standing charge—NG	29.0	p/day,th	[63]
Grid thermal equivalent	2.5	kWh,th/kWh,el	Assuming grid efficiency 40% [66]
B—Emissions Factors
Parameter	Value	Unit	Notes
CO2 intensity—Scottish grid	26.9	g·CO2/kWh,el	[63]
CO2 intensity—UK grid	153.6	g·CO2/kWh,el	[63]
CO2 intensity—vGF	1.9	kg·CO2/kg·vGF	Average of 10 reliable data points in [67]
CO2 intensity—polyester	7.6	kg·CO2/kg·polyester	[68]
Embodied thermal footprint—vGF	22.5	MJ,th/kg·vGF	Average of values reported in [69]
C—Material & Waste Costs
Parameter	Value	Unit	Notes
Selling price—vGF	1.7	£/kg·vGF	[64]
Selling price—rGF	1.36	£/kg·vGF	Assuming 80% of the vGF price
Selling price—polyester	3.0	£/kg·polyester	[68]
WTB landfilling gate fee	150.0	£/t	[33]
WTB shredding cost	580.0	£/t	From 5–10 m chunks to 1–10 mm flakes [33]
D—Capital & Operating Costs
Parameter	Value	Unit	Notes
Total CapEx	1071.4	k£	Incl. 20% contingency to consider CHP costs
Labour cost	131.4	k£/y	1 person/shift, 3 shifts, labour fee of 15.0 £/h
Maintenance	64.3	k£/y	6% of CapEx [64]
Depreciation	107.1	k£/y	10% of CapEx [64]
Overheads	113.4	k£/y	25% of OpEx incl. energy, labour, and maintenance [64]
Total SG&A	236.0	k£/y	35% of OpEx excl. shredding costs [64]
E—Plant Operating Parameters
Parameter	Value	Unit	Notes
WTB throughput capacity	365.0	t/y	The pilot
Annual operational hours	8760	h/y
GF content of WTB feedstock	67.7	wt%	[33]
E—Results of the Produced Recycled Glass Fibres
Parameter	Value	Unit	Notes
rGF production capacity	247.1	[t/y]	Considering 67.7% of WTB GF, and the rest epoxy
Embodied thermal footprint—rGF	15.03	MJth/kg rGF
Embodied electric footprint—rGF	6.49	MJel/kg rGF
Embodied equivalent thermal footprint—rGF	31.25	MJth,eq/kg rGF	By converting the electric footprint to a thermal footprint
CO2 intensity—rGF	2.24	kg CO2/kg rGF	For the plant’s location in Scotland
OpEx	1121.9	[k£/y]	Total OpEx incl. energy, labour, and maintenance, depreciation, overheads, shredding costs, and SG&A
Gross cost—rGF	4.54	£/kg rGF	Total OpEx, excl. revenues from selling rGF and gate fee
Net cost—rGF	2.96	£/kg rGF4	Total OpEx, incl. revenues

2.2. The ORC System

The ORC system is a non-regenerative, small-scale unit with a nominal electrical output of 20 kW_el and a thermal output of 130 kW_th. The selected working fluid is toluene, based on its suitability for the design temperature range of the waste heat stream [70,71], and its favourable environmental properties. The ORC system is modelled using the assumed input characteristics presented in

Table 2. The characteristics of the ORC main components and the model assumptions adopted from the mean values from the literature reported in [72].

Turbine isentropic efficiency	0.80	Superheating degree	10 K
Pump isentropic efficiency	0.60	Subcooling degree	5 K
Pump & turbine mechanical efficiency	0.95	Evaporator temperature approach	60 K
Generator efficiency	0.95	Temperature pinch	5 K
Nominal ORC net electric efficiency	0.15	Working fluid	Toluene [70,71]

. Box models are employed for the heat exchangers, with temperature approaches used to determine the system’s high and low pressures in the evaporator and condenser, respectively [72]. The temperature approach is defined as the difference between the inlet temperature of the hot or cold stream and the saturation temperatures in the evaporator or condenser, respectively. The temperature approach in the evaporator is held fixed, thereby fixing the high-side pressure of the ORC system, whereas that of the condenser is calculated as presented in the Appendix A. A fixed temperature pinch of 5 K, as adopted in [73], is used as the control parameter to represent the thermal resistance in the heat exchangers [72]. It corresponds to the minimum temperature difference between the hot and cold streams within the heat exchangers; a positive value ensures thermodynamic feasibility of the heat exchange process. The remaining two assumptions for defining the system boundaries are the fixed degrees of superheating and subcooling in the evaporator and condenser, respectively.

The pump and turbine are modelled using single-coefficient models. The combination of single-coefficient machine models with box models (temperature approach) for the heat exchangers—used to determine the high and low pressures of the system—is well suited to analyses in which the monthly and annual system performance are of interest [72,74]. This approach represents a reasonable compromise between computational cost and modelling effort, and the accuracy of the calculated monthly and annual electrical and thermal outputs. Further details regarding the ORC modelling and solution algorithm are presented in Appendix A.

2.3. The Residential Users

A five-storey building with a 100 m² floor area, with walls directed toward the four cardinal directions, is modelled using TRNBuild^®, (version 18.06.0002), with an indoor comfort temperature of 20–24 °C specified as the setpoint. The model yields the transient thermal (heating/cooling) demand of the building at a 5 min time step, based on a detailed building model. The inputs include the longitude, latitude, and the prevailing weather conditions of Aberdeen, such as solar irradiation, wind speed and direction, dry and wet bulb temperatures, and humidity. The building model requires several inputs, such as the number of floors, floor area, ceiling height, wall orientation, wall materials, window-to-wall ratio, window types, and the specification of heating/cooling equipment. The cooling demand is found to be negligible and is therefore disregarded owing to the predominantly cold weather of Aberdeen, Scotland. The domestic hot water (HW) demand is calculated in a separate TRNSYS [75] simulation in accordance with the European standard UNI EN 15316-3 [76]. The building and HW models in TRNSYS were previously presented and discussed in [4], where the occupants were coupled with a hybrid solar-wind ORC system. Both space heating (SH) and HW demands have been considered with a unitary contemporaneity factor.

The hourly electrical demand is taken from a survey of seasonal variations in the electrical demand of dwellings in the UK on working days, based on electrical appliances only [77]. The differences in seasonal demands are attributed to higher active occupancy in winter compared to summer, and the variation in daylight hours between seasons, which is more pronounced at higher latitudes. For the present analysis, the same hourly demand pattern of electrical demand is considered for each season. The data were reported for a survey of 1000 two-person dwellings [77]. Hence, a coincidence factor is introduced to estimate the electrical demand of one dwelling. The coincidence factor accounts for the asynchronous nature of the aggregate demand from all users in a network. For n apartments, it is defined as follows:

$$c={\displaystyle \frac{W\left(n\right)}{n·W_i}}$$

(1)

where $c$ is the coincidence factor, $W\left(n\right)$ is the power demand of the network of n users, and $W_i$ is the power demand of each user. The coincidence factor is estimated using a model based on empirical studies [78]:

$$c=0.5\left(1+{\displaystyle \frac{5}{2n+3}}\right)$$

(2)

The coincidence factor is then applied to calculate the electrical demand of an arbitrary number of dwellings in the simulation.

The daily total thermal energy demand of one building (five apartments) and the mean daily ambient temperature of Aberdeen are shown in Figure 2 for an entire year. The hourly electrical demand in each season is also shown in Figure 2 for a two-person dwelling, considering the calculated coincidence factor.

2.4. The Water Heating/Cooling Unit

The water heating/cooling unit consists of the water tank, the DC, the boiler, the water pump, and three control valves for circuit regulation. These are required to control the water tank temperature, provide HW and SH reliably and at the required comfort conditions, and maintain the heat sink for ORC operation of the ORC system. The characteristics of the main components of the water heating/cooling unit used in the model are reported in

Table 3. The characteristics of the main components of the water heating/cooling unit.

DC capacity ratio	0.045 [kWel/kWth]	OHTC of tank losses	1 [W·m−2·K−1]
Tank diameter	2.5 [m]	OHTC of the spiral tube HE	665 [W·m−2·K−1]
Water tank volume	700 [litre/apartment]	The diameter of the spiral tube	0.8 × tank diameter
Boiler thermal efficiency	0.95	Height of the spiral tube	Tank height/3
Water pump isentropic efficiency	0.70	The pitch of the spiral tube	4.1 [cm]
Tex,water,DC–Tamb	8 K	Spiral tube diameter	3 [cm]

. The DC and boiler are modelled using a constant capacity ratio and thermal efficiency, respectively. The DC capacity ratio is taken from the average seasonal value reported in the German standard for calculating the energy consumption of buildings [79].

Furthermore, the water outlet temperature of the DC is assumed to be 8 K above the ambient temperature, based on the specifications of commercial dry coolers. A constant OHTC is used to calculate the thermal losses of the water tank to the surroundings and the heat transfer between the cooling water and the water in the tank via the internal spiral tube HE. The geometry of the spiral tube, required to calculate its heat transfer area, is reported in Table 3. The heat transfer between the cooling water and the water in the tank is modelled using the finite volume method to capture the temperature profile of the water along the height of the spiral tube. This 1D model leads to a more accurate estimate of the cooling water outlet temperature compared to a uniform tank model.

The control logic of the water heating/cooling unit is reported in

Table 4. The control logic of the water heating/cooling unit.

Rule	Conditions	Actions
1	$T_{tk,min}<T_{tk}<T_{tk,max}$ AND Boiler and DC are off	ORC cooling water passes through the tank. Keep/turn the boiler and DC off.
2	$T_{tk}\le T_{tk,min}$	ORC cooling water passes through the tank. Turn the boiler on.
3	$T_{tk}\ge T_{tk,max}$	ORC cooling water passes through the DC. Turn the DC on.
4	$T_{tk}\le T_{tk,min}$ + 5 AND Boiler is on	ORC cooling water passes through the tank. Keep the boiler on.
5	$T_{tk}\ge T_{tk,max}$ − 5 AND DC is on	ORC cooling water passes through the DC. Keep the DC on.

. The control logic consists of five rules to control the status of the boiler, the DC, and the control valves depending on the tank temperature at the current time step and the status of the boiler and the DC at the previous time step. A hysteresis of 5 K is incorporated in the controller to avoid frequent on–off switching at short intervals. The minimum and maximum tank temperatures are set to 50 °C and 95 °C, respectively, to ensure occupant comfort and prevent boiling.

3. Results and Discussion

The results of the 3E analysis of the WTB recycling pilot (1 t/day) and the produced rGF are presented first. The primary energy consumption, CO₂ emissions, and production economics of the rGF are compared with those of vGF reported in Table 1, and the effects of WTB recycling capacity, gate fee, and rGF selling price are investigated. The pilot is upscaled using rule-of-thumb models, which provide useful estimates for the decision-making process involved in developing and upscaling a thermal recycling process. Subsequently, the life-cycle CO₂ emissions and production costs of a BMC manufacturer are compared across two scenarios: (1) using vGF and landfilling the wastes, and (2) using rGF produced by the WTB recycling pilot and recycling the wastes. In the following subsection, the performance of the proposed novel CMHP system is assessed by modelling the system over a year. The impact of the CMHP on the residential users, the WTB recycling pilot, and the overall CO₂ emissions are presented and discussed.

3.1. The 3E Performances of the rGF

In this subsection, the primary thermal energy, specific CO₂ emissions, and unit price of the rGF produced by the WTB recycling facility are compared with the corresponding vGF indices across a wide range of recycling capacities and for different gate fees. The vGF indices are drawn from the literature and reported in Table 1; however, the associated uncertainties are considerable. In particular, the range of vGF costs in [64] was 0.68–3 €/kg, and an average value of 2 €/kg was suggested for short, discontinuous fibres; this value is adopted here, noting that an average of 2.5 €/kg was reported in [80]. Regarding CO₂, a wide range of 0.6–3.6 kg CO₂/kg vGF was reported in [67] compiled from 13 databases. The mean of 10 average values of the data, with a high reliability according to the assessment in [67] is adopted in this work.

The water consumption of the GF is often neglected in the literature. It is near zero for the rGF, while it is 11.1 kg per kg vGF as the average of wet and dry chopped fibres, according to [81]. However, this figure is based on the current system design and the condition of the produced fibres being dry and without fibre post-treatment. Hence, water consumption should be considered in the sustainability assessment of GF, alongside related indicators such as water eutrophication. A more comprehensive assessment is required once the fibre post-treatment requirements are better established, following market trials on the commercialisation route.

The emissions and costs of the WTB pilot (1 t/day) are used as the baseline to estimate those of the upscaled facilities using rule-of-thumb correlations from the literature. Accordingly, the direct and indirect emissions of the process are scaled using the following empirical correlation, which estimates the primary electric and thermal energy consumption (E) of the scaled process:

$$E_{scaled}=E_{baseline}\times {\left(scalefactor\right)}^{0.7}$$

(3)

The CO₂ emissions can be calculated using the emission factor (EF) of the electricity and natural gas in the Scottish grid:

$${{CO}_2}_{scaled}=E_{scaled}\times EF$$

(4)

Regarding the costs, a similar empirical correlation is adopted:

$${\pounds }_{scaled}={\pounds }_{baseline}\times {\left(scalefactor\right)}^{0.6}$$

(5)

The exponent may vary (0.4 to 0.6) depending on the complexity of the process and system; a more conservative estimation is considered here.

The total primary thermal energy consumption of the produced rGF across scaled throughputs is shown in Figure 3 and compared with that of one of the vGF. The vGF primary thermal energy consumption is taken as the average of the reported range in [69], which was 13–32 MJ/kg and is presented in Table 1. The primary thermal energy consumption of the rGF per unit mass drops below that of vGF for WTB processing capacities somewhat higher than 1000 t/y. The specific primary thermal energy consumption of rGF is approximately 64% of that of vGF at a representative minimum commercial recycling capacity of 5000 t/y.

In Figure 4, the specific CO₂ emissions of vGF are compared to those of rGF produced in a WTB recycling facility, considering the Scottish and the average UK energy grids. The difference between the two grids is due to their different CO₂ emission intensities. In 2021, 57% of the Scottish grid was supplied by renewable sources (predominantly wind energy) and 29.8% by nuclear energy [63]. Consequently, its specific emissions are significantly lower than those of typical grids, such as the average UK grid, which had an intensity of 153.6 g CO₂/kWh_el in 2021 [63]. The CO₂ footprint of the rGF produced by the WTB recycling pilot is higher than that of vGF at pilot scale, but drops rapidly as the process is scaled up. The specific CO₂ of rGF falls below that of vGF for WTB recycling capacities exceeding 1725 t/y and 3077 t/y for Scottish and average UK grids, respectively. This demonstrates the significant environmental potential of sustainable WTB recycling processes. The adopted value in this work is the mean of 10 average values of the data with good reliability, as presented in a recent research paper [67]. However, the uncertainty of the reported CO₂ emissions of vGF in the literature is considerable. Furthermore, the rGF CO₂ footprint is higher for the average UK grid than for the Scottish grid, owing to the aforementioned differences in grid carbon intensities.

The rGF selling price relative to the vGF selling price, at which the WTB recycler reaches the break-even point, is shown in Figure 5 for a wide range of recycling capacities and different gate fees. The net revenue includes the total OpEx and the operational revenues from WTB recycling gate fees and the selling of rGF. For an assumed minimum commercial recycling capacity of 5000 t/y and a gate fee equal to the landfilling fee, the WTB recycler reaches the break-even point when the rGF selling price is 80% of that of vGF. The break-even rGF relative selling price decreases with increasing WTB recycling capacity according to an approximate power-law correlation, highlighting the importance of rGF market value in fostering a more circular economy for GRP.

The produced rGF can be used by a GRP manufacturer to replace chopped vGF in making composites. Non-exhaustive examples of GRP manufacturing processes that can incorporate rGF include bulk moulding compound (BMC), compression moulding, injection moulding of thermoplastics, extrusion, and non-woven mats. Two scenarios for manufacturing BMC with a composition of 20% GF, 20% polyester, and 60% mineral filler are compared here: (1) manufacturing BMC using vGF and landfilling end-of-life materials, and (2) manufacturing BMC using rGF and recycling end-of-life materials. In other words, the two scenarios differ in whether a recycling solution is available for the GRP, enabling BMC manufacturing with rGF and subsequent recycling, as opposed to the use of vGF and landfilling (Scenario 1). The presented cost and total CO₂ figures cover materials (GF and polyester) and BMC disposal (landfilling or recycling), excluding the cost and CO₂ of the BMC manufacturing process and mineral fillers, since these are identical for both scenarios and do not affect the comparison. The total CO₂ emission of BMC recycling is calculated based on rGF from the WTB recycling pilot (1 t/day) in Scotland, accounting for the differing polymer contents in the reactor energy balance.

The CO₂ footprint of BMC manufacturing using vGF is doubled, as presented in Equation (6), since the total CO₂ emission over the material life cycle is considered. This equation accounts for purchasing a unit mass of BMC manufactured with vGF and landfilling a unit mass of end-of-life materials; hence, the CO₂ emitted during manufacturing is included in the total emissions of Scenario 1.

$${{CO}_2}_{BMC,vGF}=2\times ({{CO}_2}_{vGF}+{{CO}_2}_{polyester})$$

(6)

Equation (7) presents the total CO₂ emission of the second scenario that the BMC is manufactured using rGF from the WTB recycling pilot, and the end-of-life of the same materials is recycled.

$${{CO}_2}_{BMC,rGF}={{CO}_2}_{rGF}+{{CO}_2}_{polyester}+{{CO}_2}_{BMCrecycling}$$

(7)

Eventually, the following equations are used for the comparison of the costs of BMC manufacturing for the two scenarios. The cost of the polyester is identical for both scenarios, but they are considered in the calculation to represent the polymer costs in the final manufacturing costs.

$${\pounds }_{BMC,vGF}={\pounds }_{vGF}+{\pounds }_{polyester}+{\pounds }_{landfilling}$$

(8)

$${\pounds }_{BMC,rGF}={\pounds }_{rGF}+{\pounds }_{polyester}+{\pounds }_{recycling}$$

(9)

The total life-cycle CO₂ emissions and the manufacturing costs of BMC are presented in Figure 6. The rGF selling price is assumed to be 80% of that of vGF, and the gate fee of BMC wastes is assumed to be the same as the WTB landfilling fee, as reported in Table 1.

The BMC manufacturer’s costs are slightly lower when purchasing rGF and recycling its products than when purchasing vGF and landfilling the waste. Regarding life-cycle CO₂ emissions, the rGF and recycling scenario yields approximately 23% lower emissions in BMC manufacturing. Therefore, a sustainable approach to BMC production using rGF and waste recycling offers both economic and environmental benefits. It is noteworthy that both rGF cost and CO₂ footprint are lower at larger recycling scales, as shown in Figure 5, further favouring the second scenario. On the other hand, if the recycling gate fee exceeds the landfilling fee, the economics of the second scenario will be adversely affected.

3.2. Performance of the CMHP System

The proposed novel CMHP system offers an alternative approach to economically valorising the WTB recycling process, complementing capacity upscaling or tariff adjustments. In addition, the integration can achieve substantial overall CO₂ reductions compared to the standalone systems. Accordingly, the results of the 3E analysis of the integrated WTB recycling pilot, ORC system, and residential buildings are presented and discussed below.

The model of the integrated system was executed with a time step of 5 min, considering varying numbers of buildings. First, a sensitivity analysis is conducted on the number of apartments integrated with the 20 kW_el/130 kW_th ORC system. The system performance is characterised by the net electric energy production ($W_{net,el}$), the net electric efficiency (${\eta }_{net,el}$), and the net actual CHP efficiency (${\eta }_{CHP}$) of the ORC-CHP system considering the delivered thermal energy ($Q_{delivered}$) to the users and the transferred heat in the ORC evaporator ($Q_{in,HF}$) as presented in Equations (3)–(5), respectively.

$$W_{net,el}=W_t-W_{p,ORC}-W_{DC}-W_{p,water}$$

(10)

$${\eta }_{net,el}={\displaystyle \frac{W_{net.el}}{Q_{in,HF}}}$$

(11)

$${\eta }_{CHP}={\displaystyle \frac{W_{net,el}+Q_{delivered}}{Q_{in,HF}}}$$

(12)

The appropriate number of users is determined to achieve near-net-zero CO₂ emissions and energy costs in a residential complex. The daily and average monthly performances of the integrated system are then presented. Finally, the economic and environmental performance of vGF and rGF is discussed from the perspectives of the GRP manufacturer and WTB recycler. The relevant economic and environmental parameters from Table 1 are considered according to each perspective. In calculating net CO₂ emissions, the specific CO₂ equivalent intensity of the Scottish electricity grid is applied. Hence, a negative CO₂ emission is attributed to the net CO₂ emissions when excess electricity is exported to the grid by the integrated CMHP system.

The comparative sensitivity analysis between the grid-supplied (Scottish grid) buildings and the buildings integrated with the CMHP is presented in Figure 7 (left) in terms of annual net CO₂ emissions and net energy costs. It is assumed that the CHP system and the buildings form a single entity, so that the buildings are the owners of the CHP system and do not pay for the produced heat and electricity, and revenues from selling excess electricity are credited to the users. This assumption serves to isolate the impact of the integration on the ongoing energy costs of the buildings.

The 30-apartment complex offers a good compromise between energy costs and the near-net-zero CO₂ emissions, as defined by the IEA [82], and a larger number of users. The 30-apartment complex requires only marginal thermal energy from the NG-supplied boiler, as shown in Figure 7 (right), while excess produced electricity is sold to the grid (positive value), resulting in near net-zero energy cost. Despite this analysis being preliminary, requiring more detailed investigation and being more relevant for larger-scale recycling plants, the results suggest that approximately 3 m² of residential area per kilogram per day of WTB recycling capacity is sufficient to achieve near net-zero CO₂ and energy costs. The use of the DC is unavoidable regardless of the number of apartments, owing to the negligible heating and cooling demand in summer in Aberdeen.

The impact of integrating the buildings with the WTB recycling pilot is presented in Figure 8 in terms of net overall CO₂ emissions. The blue line shows the annual total CO₂ emissions of the stand-alone, grid-supplied buildings and vGF production, with the annual production mass equal to that of the rGF (247.1 t/y) produced by the WTB recycling pilot. The integration of a 1 t/day WTB recycling pilot and the residential apartments lowers the overall CO₂ emissions for complexes of more than 25 apartments, and the benefit grows with a larger number of users. For fewer apartments, the CO₂ saving achieved through integration is insufficient to offset the higher CO₂ emission of rGF production compared to the vGF, resulting in higher overall CO₂ emissions. Therefore, integration enhances the environmental performance of the recycling process compared to conventional vGF production and grid-supplied buildings, even at pilot throughput.

The hourly thermal and electric performance of the integrated system with 30 apartments is presented in Figure 9 for the representative day of each of three seasons: winter, summer, and mid-season. The representative day of each month is defined as the day whose daily average ambient temperature most closely matches the monthly average temperature. The hourly results serve to assess the performance of the proposed ORC-CHP system and the implemented control logic. The figures in the left column (a, b, c) show the thermal energy flows at the water tank and its computed temperature, while the thermal energy dissipated from the tank via the DC is presented as its equivalent electric energy consumption. These figures should be examined simultaneously across different seasons to understand the interaction between the ORC and the users through the water tank, and its effect on the net electric output of the ORC system.

Examining the figures in the left column (a, b, c), the tank temperature drops to the low limit (50 °C) once in winter (January), and the boiler operates until the tank is sufficiently charged. While this indicates a good compromise between demand, supply, and storage capacity, the tank approaches its maximum charge (90 °C) in summer (August) and mid-season (May), and operates for most of the time until the tank is sufficiently discharged. A larger storage tank could be accommodated in these seasons, but this will require more boiler input in winter. Results show that increasing the number of apartments does not resolve this mismatch between supply and demand; rather, it leads to considerably more frequent boiler operation and the associated NG consumption. This is attributable to the predominantly cold weather in Aberdeen, with marginal space heating (SH) demand in mid-season and negligible space cooling (SC) demand in summer, leaving only a marginal increment in HW demand when additional users are integrated.

The thermal energy results (a, b, c) are coupled with the electric energy results (d, e, f) through the state of the water tank. The hourly electric performance shown in the right column figures reveals stable net electric production in winter, while the ORC operates frequently under off-design conditions owing to tank discharge via the DC in summer and mid-season. This leads to a sudden drop in the cooling water temperature entering the ORC condenser. The transition of the cooling water flow from the fully charged tank at 95 °C to the DC leads to a drop in the ORC’s low-side pressure, resulting in a sudden rise in the turbine pressure ratio. Consequently, the ORC system produces more net electric energy (accounting for the electric energy consumed by the DC) in summer and mid-season than in winter, and exports a greater share of excess electricity to the grid accordingly. However, a significant portion of the recovered heat from the WTB pilot is dissipated to the surroundings in summer and mid-season, despite the higher net electric output relative to winter. Therefore, the integration is overall more efficient in winter, when the thermal demand is high, as shown in the next figure.

The monthly totals of the thermal and electric performance of the integrated system with 30 apartments are shown in Figure 10a. The net monthly total electric energy production exceeds the electric demand in all months, resulting in excess electricity sold to the grid (positive values). The monthly total delivered thermal energy satisfies the thermal demand except in cold months, in which the boiler compensates for the shortfall. These effects can be seen in the electric and thermal coverage curves in Figure 10b. The net electric efficiency of the integrated system does not fluctuate significantly, with an annual average of 15.1%. This is attributable to the stable heat source at the ORC evaporator provided by the WTB pilot flue gas, and the marginal impact of off-design ORC operation on the total net electric energy production of the system.

The CHP efficiency drops significantly in summer and mid-seasons, as it is calculated based on the thermal energy delivered to the users (Equation (5)), which is considerably lower than the heat delivered to the ORC evaporator in these seasons. As shown in Figure 10a, the recovered heat by the ORC evaporator is dissipated to the surroundings in summer and mid-seasons by the DC (blue line), significantly penalising the CHP efficiency. This is consistent with the discussion of Figure 9 as well, where the hourly results highlighted the close link between the low heat demand in these seasons and the DC operating for most of the time.

The proposed CMHP system, based on the integration of the ORC system with residential buildings, demonstrates significant potential that could be further exploited in regions where cooling demand is substantial in summer. Further work is required, considering the CapEx of the ORC-CHP system, the tariffs for selling heat and electricity to consumers, a more detailed waste heat recovery configuration, and the choice between a centralised or distributed CHP arrangement.

4. Conclusions

This paper presents a 3E assessment of a wind turbine blade (WTB) recycling pilot (1 t/day) integrated with a combined heat and power (CHP) unit. This novel combined material, heat, and power (CMHP) system consists of a WTB recycling pilot, a bottoming organic Rankine cycle (ORC) system, and a water tank coupled with residential buildings. The WTB recycling pilot is configured and modelled based on experimental findings and semi-empirical models derived from an ongoing WTB recycling project. The WTB recycling pilot produces short, discontinuous recycled glass fibres (rGF) as the main product. The total manufacturing costs and CO₂ emissions of rGF production are calculated across scaled facilities and compared with those of virgin glass fibre (vGF). In addition, a BMC manufacturer is considered under two scenarios: (1) adopting vGF and landfilling the resulting GRP wastes, and (2) using rGF and recycling the resulting wastes.

The results demonstrate that the primary thermal energy consumption, CO₂ emissions, and production costs of rGF produced by the WTB recycling pilot are higher than those of vGF. However, upscaling the pilot using rule-of-thumb correlations reveals significant improvements in the 3E performance of rGF, supporting the case for large WTB recycling facilities. Moreover, further process optimisation and development can significantly reduce the carbon footprint and unit costs of rGF. Nonetheless, the BMC manufacturer realises environmental and economic benefits by adopting rGF and recycling its wastes rather than using vGF and landfilling them. These benefits are, however, sensitive to the rGF selling price, which is a critical parameter in the development of WTB recycling facilities and the establishment of a more circular economy for GF.

The CMHP system meets the electrical and thermal demands of residential users in Aberdeen, Scotland. A sensitivity analysis is performed to identify the appropriate number of apartments as the maximum count consistent with near-net-zero CO₂ emissions and energy costs. The results show that a WTB recycling pilot with 1 t/day capacity integrated with a 20 kW_el/130 kW_th ORC system is well suited to serving 30 apartments of 100 m² floor area each. Furthermore, the CMHP system achieves lower overall CO₂ emissions than the stand-alone WTB recycling pilot and grid-supplied buildings when more than 25 apartments are integrated. However, the overall CHP efficiency is low in mid-season and summer, as the recovered heat is dissipated to the surroundings via a dry cooler owing to negligible thermal demand; this limitation is not substantially mitigated by increasing the number of users, owing to the predominantly cold climate of Aberdeen.

Overall, the present analysis supports WTB recycling using fluidised-bed technology. The results indicate that large-scale recycling facilities are conducive to sustainable WTB waste management and a more circular economy for GF. A detailed model of commercial-scale recycling units will be developed in future work based on the ongoing experimental trials of the demonstration plant. Furthermore, CMHP integration is introduced as an alternative means of valorising the WTB recycling process and has demonstrated considerable potential for improving the environmental and economic performance of rGF production. Integration between large-scale WTB recycling facilities and other industrial processes that routinely require and reject substantial quantities of medium-grade heat is more promising than integration with residential users, whose thermal demand profile is highly seasonal. This will be explored in future work, considering the energy data of a relevant industrial partner.