**1. Introduction**

The current increase in global greenhouse gas (GHG) emissions should be substantially reversed to mitigate climate change and prevent global temperatures from rising beyond the 2 ◦C target [1]. Currently, global GHG emissions are attributed mainly to energy (more than 70% of GHG emissions), while the remaining are due to agriculture, land-use change and forestry, and industrial processes and waste [2]. In the European Union (EU), a binding target has been set to reduce GHG emissions by at least 40% by 2030 compared to 1990 levels [3].

In the EU, energy-intensive industries including iron, steel, cement and aluminum either have to pay carbon taxes or, if they are included in the EU's Emissions Trading System (EU ETS), they need to have sufficient allowances to cover all emissions produced [4]. Under EU ETS, installations could receive free allocation allowances every year and/or buy (remaining) emission allowances on the market. Every year, aluminum producers in the EU receive some allowances allocated for free, owing to international competitiveness of carbon leakage [5]. The number of free allocations is being decreased gradually each year, with the goal of reducing emissions and stimulating companies to invest in the transition to low-carbon technologies [4].

Aluminum (or aluminium in British spelling) is the second most used metal in the world after iron. The production of aluminum is a highly energy-intensive process, with electricity representing a large share of the energy consumed. The industry accounts for 3.5% of global electricity consumption [6]. Consequently, the aluminum industry produces a large amount of direct and especially indirect GHG emissions. During electrolysis, primary aluminum is produced via the reduction of alumina (Al2O3); this alumina is, in turn, refined from bauxite (aluminum ore). According to the data from the International Aluminum Institute [6], on average, 14.22 MWh of electricity was used to produce 1 t of primary aluminum in 2019.

Electricity can be generated from fossil, nuclear and/or renewable sources. Around half of the electricity in the EU is generated from fossil fuels, and the remaining half from nuclear power stations and renewable sources (both ≈ 25%) [7]. Wind turbines (11.4%) and hydropower plants (10.4%) account for most of the renewable energy sources [7]. Fossil fuels have been widely regarded as a prime cause of climate change because of the GHG emissions released by their burning [8]. Nuclear power plants pose serious potential risks to the environment and human health [9]. For most renewables, the main obstacles are their comparably lower energy output, intermittency and lower availability [10].

Carbon Emissions or CO2-Constrained Energy Planning is often labeled as Carbon Constrained Energy Planning (CCEP). It is a set of techniques that is suitable for power generation planning constrained by CO2 emissions [11]. Several techniques have been developed, such as insight-based approaches under the framework of Carbon Emission Pinch Analysis (CEPA) [12], algebraic targeting approaches to CCEP [13] and optimization-based targeting techniques for single- and multi- period scenarios [14]. Various extensions of the methodology have been proposed, such as an algebraic targeting approach (cascade analysis technique) for land-constrained energy planning [13], a graphical Pinch approach for water footprint-constrained energy planning, applied to biofuel production [15], improved application of CEPA to large transport systems [16], a graphical approach of CEPA applied to economic systems [17], source-sink superstructure optimization of energy planning under multiple footprint constraints [18], and various other CCEP methodology extensions. Hybrid CEPA techniques have also been developed, for instance with P-graph for macro-level [19] and plant-level planning [20] problems.

CCEP belongs to a broader category of Process Integration (PI) [21], which includes Heat Integration [22], Total Site (TS) Integration [23], Heat and Power Integration [24], Mass Integration and Resource Conservation [25], Hydrogen Pinch Analysis (PA) [26], Oxygen PA [27], Targeting Supply Chains performance [28], Targeting Property-Based Material Reuse [29], Targeting Carbon Footprint Reduction [30] and others. PI is a widely researched area, and many advancements have been made over the years. On the Science Direct platform, there are 8959 entries containing "Process Integration" in their title from 2019–2020 (search made on 8 May 2020). Also, various PI works have been published in the *Energies* journal. Among recent works are a review on progress towards efficient and clean PI [31], an analysis of PI options for different types of heat pumps through Pinch Technology [32], a proposed novel PA methodology to target cooling, heating and power in TSs [33], the incorporation of location aspects in PI methodology [34] and others, while only one work is related to carbon emissions planning [35].

Most studies applying CCEP techniques have focused on power generation, while only a few have focused on industries or specific industrial products. Tjan et al. [30] developed a graphical variant and applied it to the analysis of bulk and specialty chemicals; this method was later extended to consider allocation issues in multiproduct biorefineries [36]. Quin et al. [37] investigated product-based CCEP focused on energy-emission planning for the methanol production industry in China. Sinha and Chaturvedi [38]focused on CCEP for steel manufacturing, while Abdul Aziz et al. [39] presented a newly developed framework for low-CO2-emission industrial site planning. Recently, Sinha and Chaturvedi [40] reviewed carbon reduction technologies in industries with a focus on carbon emission limits planning.

Based on the literature review, it was found that none of the studies focused on CCEP applied to the aluminum industry and to specific aluminum products. It is also worth noting that most of the CCEP studies considered only one approach to energy planning for specific applications, while this study considers three different approaches: the first is the graphical approach of CEPA, the second is the algebraic approach implementing a cascade analysis technique, and the third is the optimization-based approach, applying a transportation model [13].

The aim of this study is the determination of the required amount of CO2 emissions from electricity sources to meet specified CO2 emission limits, considering the aforementioned approaches. All three approaches are systematic, which provides decision support in determining an appropriate energy mix to achieve a specific target, by considering the characteristics of energy sources and demands. CCEP techniques provide support in planning for more sustainable production of industrial products. They can also be used to optimize the energy supply mix based on carbon emission constraints subject to economic and other environmental constraints, except for carbon footprint and the like. This study focuses on the production of more sustainable aluminum products, and is the first CCEP work applied to the aluminum industry. CCEP approaches are applied to one aluminum product, i.e., aluminum slugs (graphical and numerical approaches), and two aluminum products, i.e., slugs and evaporator panels (all three approaches). All three approaches are finally compared and discussed.
