Facilitating India’s Deep Decarbonisation Through Sector Coupling of Electricity with Green Hydrogen and Ammonia

Abstract

Green hydrogen and ammonia are forecast to play key roles in the deep decarbonization of the global economy. Here we explore the potential of using green hydrogen and ammonia to couple the energy, agriculture, and industrial sectors with India’s national-scale electricity grid. India is an ideal test case as it currently has one of the most ambitious hydrogen programs in the world, with projected electricity demands for hydrogen and ammonia production accounting for over 1500 TWh/yr or nearly 25% of India’s total electricity demand by 2050. We model the ambitious deep decarbonization of India’s electricity grid and half of its steel and fertilizer industries by 2050. We uncover modest risks for India from such a strategy, with many benefits and opportunities. Our analysis suggests that a renewables-based energy system coupled with ammonia off-take sectors has the potential to dramatically reduce India’s greenhouse emissions, reduce requirements for expensive long-duration energy storage or firm generating capacity, reduce the curtailment of renewable energy, provide valuable short-duration and long-duration load-shifting and system resilience to inter-annual weather variations, and replace tens of billions of USD in ammonia and fuel imports each year. All this while potentially powering new multi-billion USD green steel and maritime fuel export industries. The key risk for India in relation to such a strategy lies in the potential for higher costs and reduced benefits if the rest of the world does not match their ambitious investment in renewables, electrolyzers, and clean storage technologies. We show that such a pessimistic outcome could result in the costs of green hydrogen and ammonia staying high for India through 2050, although still within the range of their gray counterparts. If on the other hand, renewable and storage costs continue to decline further with continued global deployment, all the above benefits could be achieved with a reduced levelized cost of hydrogen and ammonia (10–25%), potentially with a modest reduction in total energy system costs (5%). Such an outcome would have profound global implications given India’s central role in the future global energy economy, establishing India’s global leadership in green shipping fuel, agriculture, and steel, while creating an affordable, sustainable, and secure domestic energy supply.

1. Introduction

Carbon-free green hydrogen and green ammonia have been recognized as essential technologies for achieving net-zero greenhouse gas emissions [1]. These energy-dense fuels are expected to play crucial roles in decarbonizing heavy-duty transport (i.e., aviation, and shipping [2]), decarbonizing industrial processes such as fertilizer and steel production, providing long-duration storage and dispatchable electricity generation, and facilitating energy trade between regions [3,4]. Net-zero targets have been set by over 130 countries to date [5], and hydrogen specific roadmaps, strategies, and policies have been announced in many countries, including India, the United States of America, United Kingdom, Germany, France, Portugal, Spain, Denmark, Netherlands, Australia, Chile, Japan, and South Korea [6]. Some of these targets make reference to specific capacity ambitions, including producing 10 million tons domestically and importing 10 million tons of renewable hydrogen for the EU by 2030 [7], and 5 million tons of green hydrogen produced in India by 2030 [8].

Despite such clear signals, the relationship between new giga-scale fleets of electrolyzers and the grid has yet to be sufficiently explored in the literature using appropriate energy system modeling tools. At a global level, the International Energy Agency (IEA) predicts up to 140 million tonnes in hydrogen production capacity by 2030 [9], while the International Renewable Energy Agency (IRENA) forecasts that the production of green hydrogen and its derivatives (mostly green ammonia) will account for roughly 30% of the global electricity demand in 2050 [10]. Research using Energy System Models (ESMs), the dominant tool for understanding different scenarios of energy decarbonization, are increasingly incorporating the dynamic coupling of green hydrogen with industrial demands in net-zero decarbonized energy systems. Several ESM-based analyses have investigated sector coupling of hydrogen for various sectors, such as aviation and shipping fuel, industry, and heating [11,12,13,14]. They have also been used to evaluate flexible electrification pathways using electrolyzers [15] and evaluating hydrogen infrastructure possibilities [16,17].

While the role and potential of hydrogen in decarbonizing energy systems has been more widely recognized and researched, targets and investments are also starting to focus on ammonia and its unique role in mitigating significant CO₂ emissions. Ammonia is increasingly seen as an economically viable vector for hydrogen, necessary for overcoming the storage and transportation challenges associated with hydrogen [18]. Green ammonia production relies on historical mature production, storage, transport, and cracking technologies. Fundamentally, the Haber–Bosch synthesis process can remain largely unchanged in the fossil-based design versus the green process, although more flexibility in the ramping and minimum load level of the process become beneficial in high renewable energy systems to economically match the variability of renewable electricity and reduce costs [19].

Decarbonizing ammonia is also essential for the existing uses of the circa 180 million tonnes of ammonia that is currently produced, transported and consumed every year, predominantly for nitrogenous fertilizer. Nitrogenous fertilizer is key for feeding the global population, and although there is important research ongoing to improve fertilizer application efficiency and to reduce inorganic fertilizer applications, the long-term forecast to 2050 is still a growing ammonia demand, ranging from 123% to 137% compared to today’s levels to produce fertilizers for a growing population [6]. Secondly, low-carbon ammonia is the leading candidate for replacing fossil fuel-based maritime shipping fuels [20]. The International Energy Agency’s most recent estimates are that at least 45% of shipping fuel will be ammonia by 2050 in their net-zero scenario [6], while other projections by institutions like Lloyd’s Register are up to 100% of shipping fuel by 2050 [21]. Considering these two sectors alone, green ammonia has a mitigation potential of circa 5% of global greenhouse gas (GHG) emissions split between fertilizers (c. 1.8% of GHG emissions [22]) and shipping fuel (c. 2.9% of GHG emissions [23]). This is in agreement with IRENA’s estimate that hydrogen and its derivatives can mitigate 10% of CO₂ emissions [10].

The final category of significant GHG mitigation potential for ammonia is via the storage and trade of energy and hydrogen. IRENA estimates that half of all international trade of hydrogen in 2050 will occur as ammonia [10]. This stored or traded ammonia may be used as a fuel for dispatchable power [24] or cracked and used to deliver hydrogen to different end users [10]. The emissions reduction of this end use case is more difficult to estimate, and indeed, the outcomes of this analysis shine a light on some of this potential.

The ESMs in the literature that have included power generation and sector-coupling for hydrogen or ammonia have generally focused on short-duration, intra-daily load shifting, predominantly in the light-duty vehicle transport sector and the thermal sector, such as coordinated charging of battery electric vehicle (BEV) fleets [25], flexible heating and cooling demands [26,27], and agricultural irrigation pumping [28]. He et al. [29] begin to capture the load-shifting potential of the hydrogen sector in an ESM-based analysis of the northeast USA grid. However, this work only considered a flat weekly hydrogen demand for transport in light- and heavy-duty fuel cell electric vehicles (FCEVs), omitting other key sectors such as chemicals, steel, and aviation and shipping fuels. Given the high costs of storing hydrogen above-ground, the sector coupling explored by He et al. only results in load-shifting benefits on an intra-daily level [29].

ESM studies that are based on renewable-dominated systems have also increasingly identified the need for long-duration storage [30] or firm generating capacity [31] for achieving zero emissions in the electricity grid. It therefore becomes important to use sector coupled ESMs to explore the techno-economic potentials of technologies and energy infrastructures that offer intra-daily, weekly, seasonal, and interannual load shifting, peak generation, and sector coupling, or Power-to-X (PtX), where X represents a fuel or commodity such as hydrogen or ammonia. Previous ESM analyses have focused on synthetic methane and hydrogen gas turbines, and more recently on methanol with carbon cycling [32], to provide long-duration storage and generation flexibility. However, ammonia-fired gas turbines, that are potentially more cost effective than other clean firm power generation at low capacity factors [19], can provide similar opportunities, particularly in high renewable systems.

More recent analyses are thus beginning to highlight the role of hydrogen- and ammonia-based power generation in future electricity systems [24,33,34]. However, by not including the role of load shifting from these fuels in a sector-coupled ESM, the requirement for long-duration storage is potentially over-estimated. For example, Cole et al. find a least-cost, 100% decarbonized USA grid required over 500 GW of green hydrogen-based electricity generation via gas turbines [33]. However, Cole et al. do not consider grid-connected electrolysis for any sector coupling load shifting—not even for the hydrogen consumed in the model itself.

The only other ESM that has considered ammonia-to-power is that of Ikäheimo et al. [35]. Ikäheimo et al. [35] include ammonia-to-power via gas turbines as an option in their ESM of Northern Europe to 2050, but they do not offer a detailed justification of the capital costs or efficiency of their ammonia-to-power assumptions. Additionally, Ref. [35] does not report or compare ammonia-to-power on an LCOE basis with any competing technologies. Finally, Ikäheimo et al. [35] do not consider the ammonia synthesis flexibility, and they thus miss the holistic sector coupling potential.

A number of studies have examined the cost effectiveness and optimal sizing of hydrogen and ammonia energy production and storage in islanded energy systems [36,37,38,39]. Others have assumed islanded hydrogen is the most cost effective option at a national scale, without exploring the alternatives [40]. However, to our knowledge, no ESM study of a national energy system has compared the cost effectiveness of grid-connected and islanded hydrogen/ammonia capacity for energy storage or sector coupling. It is our conjecture that only by connecting hydrogen/ammonia production to a large energy system can a valid comparison be made between these two alternatives in terms of the overall energy system costs.

A detailed overview of the relevant national energy modeling literature for India is provided in the Supplementary Materials (Section S1), which further supports our assertion that this study addresses three clear research gaps in the ESM literature. Firstly, we include the role of flexible production and storage of both hydrogen and ammonia for intra-daily, weekly, seasonal, and interannual load shifting, where ammonia can also be used in an ammonia gas turbine for peak power generation. Secondly, we couple hydrogen/ammonia production for heat, transport, and industry to the power sector, which is overlooked in most of the other models. Finally, we compare and contrast the costs associated with grid-connected versus islanded hydrogen/ammonia on a simulated national energy system.

To accomplish this, we build upon a state-of-the-art ESM framework to explore the full sector coupling potential of hydrogen and ammonia for both give-and-take services to the grid, i.e., load-shifting and dispatchable power generation. To adequately represent the techno-economic potentials, we include multiple-weather years to consider interannual variations and analyze the system resilience. In addition, we incorporate technological forecasts based on learning curves from Way et al. [41]. This ESM framework therefore enables a state-of-the art assessment of deep decarbonization through the sector coupling of electricity, green hydrogen, and ammonia.

Using this framework, we build a national ESM of India’s electricity grid toward 2050 to explore the dynamic role of significant green hydrogen and ammonia production in a rapidly decarbonized electricity system.

India’s Energy System and Visions

Our study focuses on India due to its globally unmatched plans for growth in all three relevant sectors as follows: green electricity, green hydrogen (for steel and transport demands), and green ammonia (for fertilizer and shipping fuel demands). The ambitious decarbonization pathway we model is compiled from sectoral level studies and forecasts that suggest over 25% of the electricity generated in India is to be used for producing green hydrogen and ammonia by 2050, as shown in Figure 1. Although, such ambitious plans also require unprecedented growth in the renewable capacity of India, they are aligned with the IRENA global forecast of 30% of the total electricity demand in 2050 being used for the production of green hydrogen and its derivatives [10].

India is likely to be at the heart of any global hydrogen and ammonia transition. It has unmatched proposed growth in demand for hydrogen and ammonia, and it has already established pioneering policies to rapidly move into deploying the necessary technology. In 2021, the Indian government announced the target of achieving net-zero by 2070 [42]. Central to this plan is the announced National Hydrogen Mission (NHM) designed to accelerate the deployment of hydrogen technologies and to establish India as a global manufacturing hub for electrolyzers and fuel cells through green hydrogen obligations in the industrial production of materials such as fertilizers, steel, and petrochemicals [8,43]. The proposed green ammonia obligations in the fertilizer sector alone would cause the world’s fastest national green ammonia build out.

This front-runner positioning of India has strong economic and political motivations. Today, India is one of the world’s largest importers of fossil fuel based ammonia, with ammonia imports equivalent to 1.3 billion USD [10]. The other top importers are the United States (USD 1.46 B), South Korea (USD 746 M), Morocco (USD 589 M), and Belgium (USD 412 M) [44]. Moreover, over half of India’s imports of Liquefied Natural Gas (LNG) are used for the domestic production of ammonia; thus, India is almost completely dependent on imported fuels for its ammonia industry [45].

The proposed transformation of the Indian energy system extends well beyond producing green ammonia for fertilizer, and thus creates an ideal case study for the cross-sector role of hydrogen and ammonia in decarbonizing a national energy system. The Indian economy is poised for rapid economic growth, rapid urbanization, rapid industrialization, and rapid decarbonization—all at the same time. In global models from the International Energy Agency (IEA), the currently most populous country on the planet [46] is expected to have the largest increase in energy demand and near-term CO₂ emissions of any country [47]. Unsurprisingly, the IEA has firmly stated, “Whichever way the global energy economy evolves from here, India will be firmly at its centre” (IEA, 2021) [47].

Despite the scale of the electrolyzer fleet required for India’s ambitious plans, there has not been a modeling effort that considers the dynamic role of industrial electrification and PtX sector coupling at this scale in India. There exists a large body of recent work that evaluates the role of integrating VRE into the Indian electricity system [28,40,47,48,49,50,51,52,53] (see Section S1 of the Supplementary Materials for a detailed and systematic overview of the ESM literature focusing on the Indian energy system). Many of these ESM analyze the development of new generation, transmission, and storage assets, as well as the best way to utilize existing assets. However, all of these studies overlook the significant role of PtX sector coupling on facilitating the integration of high levels of VRE, reducing the cost of decarbonization, and reducing the need for long-duration storage, leaving a noteworthy gap in the literature given the scale of India’s ambitions.

Section 2 presents an overview of the methodology and data that are applied in this study to model the sector-coupling of green hydrogen and ammonia for the future national-scale electricity grid of India. Section 3 provides an analysis of the results and Section 4 discusses the implications of the results with some concluding remarks.

2. Materials and Methods

In this section, we present the developed methodology, main input data, as well as descriptions of the scenarios used in this study to model India’s deep decarbonization pathways.

2.1. Beyond State-of-the-Art Energy System Modeling

ESMs are a key tool for providing insights into energy system evolution for research institutions, policy makers, and energy companies. However, such long-term process-based energy system model simulations are generally very complex and difficult to validate [54]. While we have taken the recommended steps to ensure our model reflects the energy system being modeled (including running the model from a past year (2020), and verifying the actual steps to ensure the model processes are sufficiently aligned to reproduce the current levels), the uncertainty associated with such future planning models is very large. External disruptive events, like new technological breakthroughs and an increase in climate variability due to climate change, can have a major impact on the future but are rarely represented in simulations (such as changes to wind resources and changes in demand due to increased temperatures).

As summarized in Figure 2, we apply several novel methodological approaches to address these limitations and to better capture the risks and opportunities associated with the transformational changes proposed for India’s energy system [55]. In particular, we expand on previously published results from ESMs to include (1) a detailed dynamic integration of PtX sector coupling (considering production, storage, transportation, industrial use, and peak power generation), (2) a representation of system resilience to interannual weather variation at high VRE penetration, and (3) empirically grounded technology cost forecasts.

The first innovation of our work is the application of a network model of PtX based on OptiFlow modeling from [56,57], adapted to hydrogen and ammonia production, storage, transport, and use, including in the power sector itself.

The second innovation for this study deals with an important source of uncertainty, particularly relevant to a renewables-based energy system, namely, system resilience to interannual weather variation. To address this uncertainty, we make use of 10 years of hourly wind and solar data rather than the typical meteorological year used in other modeling efforts. Dowling et al. [30] find that including more years of weather data dramatically increases the need for long-duration storage in ESM results due to interannual variation, making it a relevant additional analysis for this study.

The final innovation is in the use of three different projections of global technological progress to capture the risks and opportunities for India, associated with technological progress in key clean energy technologies (such as solar PV, wind, batteries, and electrolyzers)—one of the most important sources of uncertainty in the energy transition [58], which can profoundly impact the choice of the least-cost transition pathways [59]. For this purpose, we use empirically grounded technology cost forecasts (experience curves), based on the global energy transition scenarios of Way et al. [41]. Experience curves are a reliable forecasting tool for technology progress [60], and they have been found to be significantly more reliable than the expert forecasts commonly used in modeling energy transition [61].

In this study, we consider three global scenarios for the speed of the energy transition (Historical Mix, Slow, and Fast) from Way et al. [41] (Figure 3). Through the application of these three diverse scenarios, we are able to examine the risks and opportunities for India associated with technological progress. The Historical Mix scenario can be loosely interpreted as embodying the risk that India’s ambitious deployment of key green energy technologies is not matched by the rest of the world—a scenario in which further cost declines in renewables are only modest. The Slow and Fast transition scenarios on the other hand, represent greater global ambition and hence greater global cost declines, with the fast transition representing a Paris-aligned pathway to a renewables-dominated global energy system [41].

Leading ESMs, such as those from the IEA [1,62], use cost estimates that have been shown to have systematic bias against progress in key clean energy technologies [58], with recent projections generally falling between the Historical Mix and Slow scenarios applied in our study (Figure 3). Reliable and coherent cost forecasts are essential for understanding and predicting the likely cost of India’s energy system transition. For example, Lu et al. [53] forecast Indian solar capital costs to be from USD 550 to USD 1650 kW⁻¹ by 2040, while the costs in 2020 were already USD 596 kW⁻¹ and falling [63]. Unsurprisingly, Lu et al. design an Indian electricity system which is dominated by onshore wind, rather than solar PV (amongst other differences with our study).

2.2. Energy System Model Setup and Scenarios

2.2.1. Balmorel—A Comprehensive Long-Term Energy System Model

The model we employ for this research is bespoke, built using the open source energy system modeling framework Balmorel [64]. Balmorel is designed for the exploration of total power system costs. It is demand driven and computes the conversion of primary energy to energy carriers in the form of electricity, whilst simultaneously optimizing investments and/or operational decisions, i.e., subject to policy and environmental restrictions. In its basic configuration, the Balmorel modeling framework linearly optimizes investment in generation and transmission using hourly dispatch simulation and multi-year scenario development. Essentially, Balmorel finds the least-cost economical dispatch and capacity expansion solution for the represented energy system, subject to any technical and economic assumptions and constraints provided.

The Balmorel model built for this study covers a large geographical area of the Indian power system, whilst allowing for spatial analysis focusing on, for example, the production of green hydrogen and ammonia. We divide India into the historical five electricity grid regions. Each region consists of one or more areas, which represent the PtX networks or installations of local electricity technologies, with electricity and fuels traded between adjacent market regions. The temporal resolution in Balmorel is defined by the user, allowing the system to be simulated using an hourly time resolution or to be aggregated according to the research question. The model is partial equilibrium, meaning that it considers the electricity sector and components of the coupled sectors of the economy (including the feedback between these sectors), but it does not model the full economy. It is a deterministic model which assumes perfect market competition and economically rational decision makers seeking least-cost solutions.

The decision to use Balmorel for this study is detailed in the Supplementary Materials (Section S2), but it primarily revolves around its ability to scale from regional to international power systems, and that it is customizable and open source. Balmorel has been applied extensively around the world since its release in 2001, including similar national energy transition capacity analyses in China [65], Indonesia [66], and Vietnam [67]. The mathematical formulation and results of Balmorel have been compared and validated against other recognized energy system models [68,69].

Balmorel has previously been used to analyze scenarios with future hydrogen demands, focusing on European hydrogen infrastructures [17], the potential role of offshore wind in the North Sea [70], the potential use of Danish thermal heat storage under hydrogen export scenarios [71], and for renewable gas utilization [72]. In addition, Balmorel has been used to model green ammonia for fertilizer and dispatchable power generation in Northern Europe to 2050 [35]. However, the latter analysis does not consider flexible Haber–Bosch (HB) ammonia synthesis plants, and thus, no meaningful load shifting can occur. Additionally, other hydrogen demands are not considered, such as steel or maritime fuel. Nevertheless, the results of Ikaheimo et al. [35] pointed towards the potential use of green ammonia for energy storage and energy transport, with significant above-ground ammonia storage.

2.2.2. Optiflow—Generalized Spatio-Temporal Network Optimization Model

OptiFlow [73] is an open source spatio-temporal network optimization model which can be linked with Balmorel. OptiFlow uses node–arc relationships to represent flows such as energy, mass, economic, or environmental metrics. Nodes include storage, transport, and chemical processes, such as ammonia synthesis. Arcs in this model include electricity, hydrogen, and ammonia. In the linkage with Balmorel, the OptiFlow objective equation of minimizing cost is included inside Balmorel’s cost optimization objective function. Other linked equations include the electricity balance equations and ammonia balance for re-electrification in gas turbines. In this study, green hydrogen and ammonia production, storage, transport and use are modeled using OptiFlow linked with Balmorel, based on the configuration used in [56,57] (see Bramstoft et al. [56] for a detailed description and equations in OptiFlow).

Finally, OptiFlow is deterministic, i.e., there is no probabilistic element, and a model run will reliably produce the same result based on a fixed set of inputs.

2.2.3. Scenarios

In this study, we investigate six national scenarios encompassing three different speeds of global transition (Historical Mix, Slow, and Fast) across two different network configurations (connected and islanded). The Historical Mix scenario assumes a very slow uptake of new renewable technology, and hence cost declines, basically keeping the mix of renewables in the global energy system constant at the current levels (around 10%). The Fast Transition assumes renewable deployment grows at current rates for another decade and then follows a standard S-curve to slow to around 2% once these technologies become dominant. In the Slow Transition, the current rapid deployment trends for renewables slows down immediately but continues to grow at lower rates for a decade, such that global deployment lies between the Fast and Historical Mix scenarios. The detailed techno-economic assumptions used for the application of these scenarios to this study are provided in the Supplementary Material (Section S4).

2.2.4. Spatial and Temporal Aggregation

We consider 35 areas (Indian States and Union Territories excluding the small Andaman and Nicobar Islands and Lakshadweep islands), which are combined to form 5 Regions (East, North, South, West, and Northeast). Unlimited transmission is assumed within regions and limited transmission between regions based on the existing and new infrastructure. Data for the existing and planned generation and transmission were used for each area based on the National Electricity Plan (NEP) [74] and Electric Power Survey (EPS) [75].

Although our system model is complex, it contains some necessary simplifications, including a simple representation of electricity transmission (a single capacity between regions), short-sighted investment optimization (i.e., no inter-year foresight to anticipate falling technology prices or rising CO₂ prices), full intra-year foresight (e.g., hydro use, storage, and fuel restrictions using perfect prediction of wind and solar), and perfect competition in the energy market. We modeled the development of the system to 2050 in 10-year time steps. We use ERA5 [76] climate data for generating the variability of hourly solar and wind profiles across the whole period of 2010–2019 (see Supplementary Materials Section S4). We also aggregate every 5 h and every 5 weeks over 522 weeks (10 years) to capture the hourly weather variation. This allowed us to model at a fine enough timescale to capture renewable variability, combined with interannual weather variations, without causing untenable calculation times. Each of the model runs took between 3 and 72 h on a 256 GB RAM, 2.2 GHz computer.

2.3. Supply-Side Technologies

Our Balmorel model considers a range of available power generation technology options including coal, gas, wind, solar, nuclear, hydropower, batteries, hydrogen/ammonia, and biomass, with India-specific cost forecasts and availability. Cost scenarios for all technologies are listed in Supplementary Tables S5–S8.

2.3.1. Technical Potential for Solar and Wind

We use 48 wind locations and 33 solar locations across India to capture the spatial variation in profiles across areas of significant technical potential. The National Institute of Wind Energy (NIWE) estimates a maximum of 693 GW of onshore wind installed capacity at 120 m based on the suitable land, divided between the states [77]. The National Renewable Energy Laboratory (NREL) RED-E tool [78] was used to identify over 9000 GW of solar PV potential capacity in India, using land classified as barren land, wasteland, and shrubland. This is the same technical potential used by The Energy and Resources Institute (TERI) [45]. See Supplementary Materials (Section S4) for further information.

2.3.2. Green Hydrogen and Ammonia Production and Storage

Green hydrogen and ammonia production was modeled using electrolyzers, above-ground hydrogen storage in tanks, flexible HB ammonia synthesis and air separation units (i.e., nitrogen generation), and above-ground ammonia storage tanks. Underground hydrogen storage is possible in certain geologies, such as salt caverns; however, India is unlikely to have many such suitable formations, and further work would be required to understand India’s geological storage options in other formations such as rock caverns [45]. Therefore, above-ground tanks were the only storage technology considered in this analysis for hydrogen.

In the past, ammonia production has typically been modeled as islanded systems [36,79], and it was only rarely modeled in grid-connected scenarios [80]. We consider both configurations, with batteries and ammonia gas turbines available in both cases for added reliability.

HB operation flexibility is a key assumption in this analysis which allows for seasonal load-shifting. HB plants in operation today do not have flexibility below 50–60% of the minimum load [81] because the source of feedstock hydrogen is fossil fuel-based, and thus the least-cost strategy is to run continuously at maximum throughput to increase asset utilization. Given the numerous green ammonia commercial announcements since 2020, HB technology companies are also announcing much lower minimum loads in greenfield plants built to connect with VRE. Patents have been filed for as low as 10% of the minimum load [81], and chemical engineering modeling in the literature suggests minimum loads of 30% can be achieved by increasing the inerts in the synthesis loop and adjusting the ratio of hydrogen to nitrogen, all while maintaining the current reactor design and auto-thermal operation [82]. This analysis assumes a 20% minimum load, which could be met by turning assets down or by designing plants with multiple smaller reactors and selectively shutting them down for periods at a time, i.e., seasonally. We use an average cost per MW to calculate the fixed capital costs. This is a simplifying assumption. Electrolyzer stacks are modular but the balance of plant costs per MW are likely to be higher for smaller plants. A profit margin is included in the capital cost calculation regardless of their utilization; however, our operating costs do not include a margin for flexibility. The other processes which demand substantial amounts of hydrogen, namely steel, refining, and transport, were not assumed to have any load-shifting capability.

No import or export of hydrogen or ammonia was considered. We assume imports will not have a significant cost advantage because India is likely to have low costs for green hydrogen and ammonia production (due to very cheap renewables and low labor costs) [3]. Exporting green hydrogen and ammonia is a strategic target announced in the NHM [43], but it is difficult to model with any certainty due to the unclear supply and demand relationships at a global level, with major uncertainties around countries’ preferences for hydrogen and ammonia fuel security, subsidies for domestic industries such as steel and fertilizer, etc.

2.3.3. Dispatchable Power Generation Requirements

Dispatchable power is the electricity generated on-demand regardless of weather conditions. This includes hydro-power equipped with large reservoirs, nuclear power, biomass power, coal or natural gas with or without CCS, and hydrogen- and ammonia-based power plants. The short-term storage of several hours in batteries, while crucial for the system, is not included in the classification of dispatchable electricity in this analysis; however, long-term chemical storage in the form of ammonia is considered dispatchable.

For the purposes of our scenarios, no new fossil fuel generation capacity additions were permitted after 2039. The existing coal plants are phased out by 2039 and the gas plants by 2049. A linear CO₂ tax from USD 45 (2025) to USD 200 (2050) was used to facilitate this phase out, in line with IEA forecasts for developing countries [62]. To replace this dispatchable power, green ammonia power generation is included using combined cycle gas turbine (CCGT) power plants [19]. Due to the high cost of above-ground hydrogen storage, ammonia has significant techno-economic advantages for use in dispatchable gas turbines [19], and there are clear signals from manufacturers to have ammonia-fired gas turbines on the market by 2024 [83]. Accordingly, we allow for new investments in ammonia CCGT from 2030, alongside other power generation technology options. Any ammonia consumed in the power sector was an additional endogenous ammonia demand (i.e., generated within the model run) on top of the exogenous ammonia demands (i.e., fixed fertilizer and maritime fuel demands).

2.4. Demand Scenario

Based on scenarios from TERI [45] and the World Bank [20], we developed a decarbonization scenario where the production of green hydrogen and ammonia could account for more than 25% of the electricity demand by 2050 in India, as shown in Figure 1. These demands total 15 MMTPA of hydrogen and 120 MMPTA of ammonia by 2050, which are significantly larger than India’s present-day fossil fuel-based hydrogen and ammonia production capacities of 6 MMTPA and 18 MMTPA, respectively. The modeling underpinning this low-carbon demand scenario by TERI [45] assumes 50% of steel, 60% of ammonia fertilizer, and 30% refining demand are met by green hydrogen. Shipping demand is based on supplying 10% of global shipping fuel (meeting 25% of Asia’s fuel demand) based on the most conservative scenario for India from [20]. Indian electricity demand is modeled by TERI [45], based on the detailed modeling of electricity growth in the residential, commercial, transport, industry, and agricultural sectors. Further data are available in the Supplementary Materials (Section S4).

Exogenous demands for hydrogen and ammonia do not include long-duration storage needs (which are determined endogenously) or synthetic aviation fuel, which is still technologically uncertain. No demand-side flexibility in residential heating or vehicle charging was considered; however, these intra-daily load-shifting technologies would only affect the deployment of 4-h battery storage, which is not within the scope of this research focusing on seasonal load-shifting and storage requirements.

3. Results

This section investigates the simulation results for India’s energy system broken down by electricity network effects, seasonal storage and system resilience requirements, and their dependency on the speed of renewable generation deployment.

3.1. Network Effects: Why Connect Power-to-X to the Electricity Grid?

The results of our modeling provide evidence of significant benefits for India from connecting hydrogen and ammonia production to the electricity grid, compared to relying on predominantly islanded production sites. In our simulations, grid integration reduces the Levelized Cost of Hydrogen (LCOH) and Ammonia (LCOA) by 10–25% across each of the scenarios (Figure 4a,b). The reduction in LCOH and LCOA is driven primarily by the lower cost electricity available to the grid-connected electrolyzers, cheap electricity that would otherwise have been curtailed. In the islanded production scenarios, 15–20% of electricity generated in India in 2050 is curtailed, while only 10–14% is curtailed in the grid-connected scenarios (Figure 4c). In absolute terms, this prevents 350 to 460 TWh of electricity from being curtailed, and consequently 200–300 GW less PV and wind capacity needs to be installed by 2050.

By reducing curtailment, connecting electrolyzers to the grid also reduces annual system costs by a modest 5% in 2050 across all scenarios (Figure 4d). Such a modest improvement is probably less than the uncertainty in the model assumptions. However, beyond reducing system costs, the grid-connected electrolyzers also enable a reduction material consumption and land-use associated with the extra solar PV that is no longer required.

As with other products with seasonal inputs, such as agricultural outputs, grid-connected electrolyzers and ammonia synthesis plants can be operated seasonally to access cheaper, otherwise curtailed electricity, with large quantities of ammonia stored to meet annual demands and interannual variations (Figure 5). Across all scenarios, ammonia production mostly occurs from March to August, with plants’ loads reduced to the minimum technical levels outside of these months (Figure 5a). These months correspond to the strongest solar resource in the Northern Region, and the strongest wind resource in the Western and Southern Regions—due to monsoons. Ammonia storage is required for the system to maintain a constant output of ammonia for fertilizer and shipping demands, ranging from 15 to 40 MMT of ammonia storage (Figure 5b).

Simulating the 2050 system with weather data from ten consecutive weather years (WY1–WY10) results in years where the system’s storage levels are nearly empty in January and reach a maximum in August. The difference in weather years (WY) can be seen in the ammonia storage levels, with WY9 and WY10 being particularly good years compared to others, generating surplus storage levels by the end of WY10 (Figure 5b).

The Historical Mix cost scenario uses more storage than the Slow or Fast scenarios because, in the Historical Mix scenario, it costs less to store a larger buffer of ammonia than overbuild solar and wind. This is due to the higher technology costs of solar PV and wind in the Historical Mix scenario. Still, all scenarios, regardless of technology costs, find seasonal ammonia production with storage beneficial for a least-cost system. While significant ammonia storage is used, no scenarios utilized ammonia or hydrogen trading between regions via pipelines, trucks, or ships for industrial demands (i.e., steel, fertilizer, transport, refining, and maritime fuel). The ammonia production cost difference between regions is calculated at less than USD 50 per metric ton, which is a relatively small difference, but enough to justify local production. Economic, social, and political differences between regions might alter such an outcome, leading to ammonia transport and ultimately higher overall costs in some regions. However, from a pure cost minimization perspective, these results point to a strategy of moving electricity via an increase in grid-connected transmission lines rather than moving energy-storing molecules like hydrogen and ammonia.

3.2. Dispatchable Electricity Generation Requirements for Seasonal Storage and System Resilience

Across all scenarios, dispatchable electricity represents from 3% to 6% of electricity consumption in 2050 (Figure 6a). Dispatchable electricity is mostly delivered via the existing and planned hydro-power and nuclear power plants in India. Nevertheless, ammonia-to-power is still used in all but the Fast connected scenario to provide seasonal storage and resilience in unfavorable weather years. Green ammonia is used to generate up to 0.4% of consumed power, or 30 TWh per year based on 70 GW of installed capacity built from 2040 onwards. These plants have an average utilization factor from 2% to 5%, far lower than the current power plant utilization factors, and a utilization range in which ammonia-based power plants are more likely to have a cost advantage [19].

In the Islanded scenarios, ammonia is used for seasonal power generation and system resilience in the Oct–Jan period (Figure 6b). The key advantage to the connected scenarios, is that the requirement for seasonal dispatchable electricity is reduced because of the demand-side flexibility gained from the ammonia production. In short, islanded scenarios use seasonal dispatchable power generation (i.e., supply-side flexibility) because they do not have the demand-side, load-shifting flexibility of grid-connected electrolyzers to manage the seasonality of the renewable resources. As shown in Figure 6b, WY1 and WY5 required the most dispatchable electricity via ammonia-to-power for providing system resilience.

3.3. Renewable Dominance and Decarbonization

Regardless of the scenario, and therefore the cost reductions in renewables, the evolution of the Indian power system depicted always tends towards one that is dominated by solar power with battery storage (Figure 7a). Across all scenarios, solar PV installed capacity reaches over 4000 GW by 2050, with 1000 GW of 4-h battery storage. The installed capacity is concentrated in the Northern Region (NR), Western Region (WR), and Southern Region (SR) (Figure 8). Onshore wind power is mostly installed in the WR and SR, totaling 330 GW to 460 GW in the country by 2050. Electrolyzer capacity reaches nearly 600 GW by 2050, representing a source for load shifting almost as large as that provided by batteries. Transmission infrastructure increases by over 60% in the connected scenarios, while only up to 20% in the Islanded scenarios (Figure 7b).

The electrification of the Indian economy increases as electricity generation rapidly shifts towards wind and solar, and it increases to solar over time as solar PV system’s historically faster learning rates begin to dominate (Figure 7c). Carbon emissions from the power system begin to rapidly decline to near zero by 2040 and achieve zero emissions by 2050 across all scenarios (Figure 7d), driven by VRE technology cost declines, a gradual CO₂ tax to USD 200 t⁻¹ by 2050, and the phase out of existing coal and gas (for power and ammonia production) by 2040 and 2050, respectively, (see Section 2 for the assumptions). No new nuclear, hydro-power, or biomass power plants are built beyond what is already under construction or planned (see Supplementary Materials for details). New ammonia-to-power plants come online in 2040, specifically in the SR, NER, and ER.

4. Discussion and Conclusions

As momentum builds with more countries transitioning towards net-zero emissions, it is important that energy system modelers continue to expand their scenarios to explore all new clean energy pathways available for deep decarbonization. Our results suggest a more foundational role of green hydrogen and ammonia should be considered as it becomes increasingly clear that fleets of electrolyzers will be deployed at a considerable scale over the coming decades [85].

In our modeling, we found that the connection of large fleets of electrolyzers into the grid infrastructure and load-shifting green ammonia production for fertilizers and shipping fuel is a plausible strategy towards a less costly, more efficient, and more reliable electricity grid in India. The opportunistic seasonal production of ammonia with storage was shown to be beneficial for a least-cost system in all the scenarios we explored. In such a least-cost system, green hydrogen and ammonia are produced in a seasonal pattern to match renewable energy surplus. Excess green ammonia is stockpiled in low-cost above-ground storage, to be consumed in the seasons of lower production, allowing for year-round demand in the fertilizer and shipping sectors to be met. Smaller quantities of green hydrogen are stored above-ground for firm generating capacity and intra-daily load shifting on the electricity grid. Such a PtX sector-coupled system configuration appears to not only be less costly but also more resilient to unfavorable variable renewable weather years, as we have shown using ten years of weather data.

The growth of demand modeled in this study is very ambitious, but our results suggest significant opportunities and benefits could ensue from such a bold strategy, at reasonably low levels of risk. By pursuing such a strategy, our model of India was able to replace approximately 50% of their domestic steel and fertilizer production with domestically produced green hydrogen and ammonia by 2050, and produce approximately 10% of the predicted global demand for green shipping fuel. Along with generating new multi-billion USD green steel and maritime fuel export opportunities, this new sector-coupled energy system has the potential to replace tens of billions of USD in ammonia, LNG, and fuel imports each year. It could also dramatically reduce India’s greenhouse emissions, provide valuable short-duration and long-duration load-shifting and system resilience to inter-annual weather variations, reduce the curtailment of renewable energy, and reduce requirements for long-duration energy storage or firm generating capacity.

Achieving such benefits requires a dramatic increase in the deployment of renewables in India, including an additional 500–1000 GW of solar PV by 2030, and at least another tripling by 2050 (Figure 7c). Such a growth in solar PV deployment is higher than China’s recent 20% per annum growth, which is the highest sustained growth of any country to date. However, such growth is technically plausible, given that planned solar manufacturing capacity is set to double by 2030 [86] and that solar PV generation is now cheaper than all other sources of electricity generation in India [62], which likely to decline more with further global deployment [60]—a particularly important source of risk that was a focus of our modeling effort.

In a scenario where the renewable costs continue to decline, in the Fast transition scenario, the levelized costs of hydrogen is already below the USD 2/kg by 2030, which is considered the threshold necessary to make green hydrogen competitive with gray (fossil fuel-based) hydrogen (Figure 4a). By 2050, green ammonia is also at the lowest end of the historical range of gray ammonia (Figure 4b).

The key risk for India of such an ambitious hydrogen strategy lies in the potential for higher costs and reduced benefits should the rest of the world reign in their deployment of renewables and storage technologies—risks that are embodied in our Historical Mix scenario. We show that this lack of global deployment, and the lack of learning-by-doing cost declines for renewables, would result in the predicted average levelized cost of green hydrogen and ammonia remaining high, although still within the historical range of their gray counterparts (Figure 4b). However, we believe the likelihood and therefore risk of such a scenario manifesting is relatively low. This Historical Mix scenario involves a rapid reversal in the current trends of global renewables and electrolyzer deployment, not just in one country such as the US, but globally. This scenario most closely matches the SSP5-RCP8.5 scenario of the IPCC in terms of its modest deployment of renewables and global emission reductions [58]. Although there is no quantified likelihood associated with any of the IPCC scenarios, the RCP8.5 scenarios are increasingly being regarded as unlikely by the scientific community, in no small part due to the lack of global deployment of renewables these scenarios project [87].

Unfortunately, the full cost benefits on the system design modeled in this research will not be realized unless policy continues to steer industry-grid relations away from some of its historical precedence. For industry, grid connection is often associated with high grid connection charges and unreliability, which warrant on-site power backup and complete “captive” power plants in extreme scenarios. This historical path dependency is extremely evident in India, where captive power plants at commercial and industry users accounted for 17% of the country’s generation in 2019–2020, with over 78 GW of captive power generation installed [88]. These captive power plants are either fractionally or completely disconnected from the distribution and transmission grids to avoid the historically under-performing reliability of the grid and expensive rate charges for industrial users.

As it stands today, electrolyzer fleets for industry in India would likely follow in the same footsteps, with islanded systems avoiding grid connection charges and other real or perceived disadvantages. However, based on our findings, policy-makers should consider steering the system towards a new paradigm of industry-grid relations that could synergistically benefit both parties in the transformation towards net-zero emissions. Recent policies announced as part of India’s National Hydrogen Mission are a step towards grid connecting PtX. One policy aims to reduce grid costs for green hydrogen and ammonia plants by waiving interstate transmission charges among other incentives [8,89]. The suitability of these policies and further policy and market mechanisms need to be explored today, at the initiation of this transformation, in order to ensure progress towards a more integrated future scenario.

Beyond India, the potential role of load shifting in green hydrogen and ammonia needs to be modeled for other countries and regions, given each region will have its own unique local VRE supply and demand mismatches, as well as specific cross sector PtX demands. Furthermore, the evolution of new technologies must be monitored, and any modeling effort adapted to best capture imminent changes, such as the cost declines in competing electrolyzer technologies, increased aviation e-fuels demands, or the discovery of geographies with large underground hydrogen storage capacity. Further work should also investigate the global trade of green ammonia to balance seasonal production [90], rather than only considering stockpiling in an isolated country. Moreover, we only consider resilience to ten years of historical weather data, which may not be sufficient considering the changing weather patterns possible under climate change. These additional resilience requirements could significantly increase the role of dispatchable power and should be considered, given climate models are able to provide these data.

There are other limitations to this modeling effort, including its reliance on simulated climate wind data for the region, not including the limited interconnections to other countries, and its simplified grid connection and distribution costs (for more details see Supplementary Materials Section S2.4). We do not believe these limitations detract from our findings—that a sector coupled grid-connected strategy is worthy of consideration by the Indian government. However, further research might be prudent to confirm our results are robust to all such limitations, as well as the fundamental uncertainties and complexities that limit our ability to predict the exact mix of our future energy systems.

Ultimately, the real value of ESMs, such as the one presented here, is not in the precise numeric results, but rather in highlighting any risks and opportunities of strategic system designs, particularly the potential for non-marginal system transformations [55]. In this spirit, we can confidently assert that hydrogen and ammonia have the potential to be leveraged to fundamentally change electricity systems for the better, if fully integrated. We hope this analysis motivates the expansion of national ESMs to also include PtX with sector coupling and the consideration of such sector-coupled energy transition strategies by policy makers.