Optimal Design of a Hybrid Energy System for Economic and Environmental Sustainability of Onshore Oil and Gas Fields

Abstract

The pollution caused by gas flaring is hazardous to the health of flora, fauna, and humans settled around the flaring site. Gas flaring also incurs economic loss as natural gas, an energy source, is wasted in flares. Furthermore, the unreliable electrical infrastructure is a roadblock for oil and gas companies attempting to achieve their production targets. This paper presents a framework to design hybrid energy systems (HES) which utilize the gas flare waste along with the locally available renewable energy sources to generate electricity. A novel dispatch strategy to suit the requirements of the oil and gas fields has been used for real-time simulations and optimization of the HES. As a test case, six different hybrid energy configurations, modelled for two gas flaring sites, Lakwa and Geleky in Assam—India, were analyzed and compared on the basis of economic and environmental factors. The best suitable configuration comprised 2000 kW solar photovoltaic (PV) panel sets, one 200 kW gas microturbine, two 30 kW gas microturbines, and grid connection. The proposed system economically outperformed the existing power system in the area by 35.52% in terms of the net present cost. Moreover, it could save 850 tons of carbon dioxide emissions annually, and it has a renewable fraction of 93.7% in the total energy generation. Owing to these merits, the presented technique would be a promising option for generation of electricity from flare gas waste and to mitigate pollution.

1. Introduction

Gas flaring poses a severe environmental threat not only in oil and gas fields but also in natural gas industries, petroleum refineries, and petrochemical plants. Apart from ecological damage, severe health abnormalities such as cardiovascular diseases are common in humans and animals living near gas flaring sites [1]. The emission levels for onshore oil and gas fields are much higher than those from other mentioned sources. Globally, more than 100 billion cubic meters of gas are flared annually in onshore oil and gas fields, which is associated with 400 million tons of carbon dioxide (CO₂) emissions per year along with oxides of nitrogen, sulfur, carbon monoxide, benzene, toluene, xylene, styrene, naphthalene, and black carbon formaldehyde emissions [2,3]. Approximately 20 million dollars are wasted due to flaring of natural gas produced at oil and gas fields across the globe. Numerous gas flaring reduction technologies have been adopted worldwide; they are mainly applied for either power generation (popularly known as gas-to-wire) or storage by liquefaction (viz.-pipeline natural gas, liquefied petroleum gas, liquefied natural gas, compressed natural gas, natural gas hydrates, gas-to-liquid, and methanol–ammonia production). Nevertheless, there is limited research on integrating the methods used for power generation (gas to wire) with the emerging renewable generation methods to limit carbon emissions due to flaring and meet the local community’s electricity demands. It can be seen that combining renewables like solar and wind introduces intermittencies and stochasticity in the system due to dependence on the dynamic weather parameters. However, studies [4,5] have proven that a mixture of multiple energy resources can effectively overcome the problems of randomness, unpredictability, and variability of solar or wind energy [6]. Hybrid (multiple source-based) energy systems significantly reduce the net present cost, increase energy efficiency, and provide a more reliable supply of electricity as compared to a single energy source system [7].

Numerous studies have been conducted on modelling, simulation, unit size optimization, and techno–economic assessment of hybrid energy systems (HES). The primary objective identified from those studies was maximizing the renewable fraction while minimizing the overall cost of the system design. For the design optimization of the hybrid renewable energy system (HRES), a number of metaheuristic optimization techniques and software are used. In [8], Mo et al. used a branch and bound method and radial basis function artificial neural network (RBF ANN) algorithm to a maximize renewable fraction at a low cost while designing an HRES using the Modelica language. The capital and cost of energy are minimized in [9] using cuckoo search and compared with a genetic algorithm and Particle Swarm Optimization (PSO). Ref. [10] compared results for PSO and a Hybrid Optimizer for Multiple Energy Resources (HOMER) and found that the system designed by HOMER provides a sufficient sizing margin while designing the system components, which leads to less load loss. The authors of [11] compare the microgrid design using HOMER and an analytical method and conclude that the HOMER designs are techno–economically more suitable as the analytical method produces exaggerated component sizes. Meta-heuristic computational approaches do not produce reasonable results when the hybrid system complexity is higher. Moreover, PSO and HOMER give similar results with less than 1% variability between them. Therefore, HOMER is more popular in the research community to reduce the computation time, improve the accuracy of results, and produce a thorough techno–commercial analysis of the production mix, and effective comparison of all possible system designs. Sen et al. [12] were among the first few to use HOMER to design an off-grid HRES comprising small-scale hydropower, solar photovoltaic systems, wind turbines, and bio-diesel generators for remote rural areas village in Chhattisgarh, India. It is to be noted that 79.34% of studies in the existing literature are purely based on ‘off-grid/islanded’ solutions [6]. In a similar grid independent study [13], Wondwossen et al. designed a wind turbine generator, diesel generator, and solar panel-based hybrid renewable energy system for a remote rural community in Ethiopia. In another grid-independent study, Murthy et al. [14] designed a stand-alone polygeneration microgrid catering to the electrical, thermal, and hydrogen needs of a typical 50-household remote Indian village. Shahzad et al. [15] designed a grid-independent PV/biomass for a remote agricultural farm in Punjab, Pakistan. Similarly, Rezk and Dousoky [16] presented a feasibility study of a stand-alone photovoltaic, a battery, a fuel-cell, wind-based HRES for a remote agricultural area in Egypt. In [17], Gonzalez et al. used small wind turbines (SWT) for off-grid electrification of 2000 kWh load in two different rural Venezuelan communities. In a similar rural context in West Africa, Ayodele et al. [11] designed an optimal hybrid configuration comprising WT, PVG, gasoline generators, and BESS for a microfinance bank. In a similar Malaysian study, wind turbine and PV-based off-grid HRES were designed by Shezan et al. [18]. Multiple researchers have explored off-grid HRES for remote rural areas as well as islands. In a feasibility study for Lakshadweep islands, Singh et al. [19] found that PVG, WT, and BESS is the best combination to lower the emission, cost of energy, and dummy load of the existing diesel operated electricity production system in the region. In a similar study, Fazelpour et al. [20] declared that a wind–diesel hybrid system with BESS is ideal for supplying power to a hotel in Kish islands, Iran. Bhakta and Mukherjee [21] studied the feasibility of a PVG–BESS powered system for a single household load in Andaman and Nicobar Island. Similarly, Tsai et al. [22], and M B Khan et al. [23] also proposed HRES over diesel generators for the Batan Island, Philippines, and a resort island in the South China sea, respectively. Later on, Tsai et al. analyzed the impact of increasing the renewable fraction on system cost for a PVG/DG/BESS system in a Chinese island [24]. Some microgrid system design studies have been conducted for academic institutions/university buildings. In [25], Kumar et al. used HOMER to design a microgrid system for an educational institution that utilizes the biogas from the institute’s food waste to generate electricity. Similarly, Sarkar et al. [26] designed and implemented a microgrid comprising solar PV, wind, biomass, vanadium redox flow battery storage for Indian Institute of Engineering science and technology (IIEST) Shibpur. Bhattacharjee et al. [27] analyzed the contribution of PV and wind in a PV–wind-based HRES for an educational building in Tripura, India. Meanwhile, Saiprasad et al. [28] performed optimal sizing and an HRES feasibility study for Victoria University at the St Albans campus in Melbourne. Ghenai et al. [29] analyzed the performance of an off-grid PV–FC–DG-based HRES using HOMER software for a university building. A few studies for urban residential buildings can also be highlighted. Liu et al. [30] performed a techno–economic feasibility analysis of hybrid renewable energy power systems for high-rise buildings in an urban location. Similarly, Al-Ammar et al. [31] designed a stand-alone HRES for a residential area and compared nine different PV/Wind/diesel/battery combinations using HOMER software to determine the most economical and environmentally safe configuration. Lv et al. [32] used HOMER to perform a techno–economic feasibility analysis of a 100% renewable energy-based residential household in China in a similar context. Tiwary et al. [33] devised a biomass-based community-scale hybrid energy management system for waste and energy management in the UK and Bulgaria urban areas. Rad et al. [34] designed a hybrid PV/wind/biogas/fuel-cell-based system for stand-alone and on-grid applications. In all these studies [11,12,13,14,15,16,17,19,20,21,22,26,28,30], there was no mention of the strategy used to schedule the dispatch of various HRES components.

Table 1. Comparison of HRES designs from the literature.

Ref	HRES Mix	Load Type	Grid Connection	Dispatch Strategy	Performance Measure
[13]	Wind, Solar, Diesel	Rural community	Off-grid	No mention	cost
[15]	Solar, Biomass	Agricultural	Off-grid	No mention
[16]	Wind, PV, Battery storage, Fuel cell	Remote Agricultural Area	Grid	No mention
[17]	Wind	Dessert and mountain	Off-grid	No Mention	LCOE
[18]	Wind, Diesel, Battery, PV	Daily load	Off-grid	CC	NPC, COE
[19]	PV, Battery	Islands	Grid	No Mention	NPC, COE
[20]	Wind, Diesel, Battery	Hotel	Grid	No Mention	COE
[21]	PV, Battery	Island	Off-grid		COE, NPC, RF
[22]	PV, Diesel, Wind, Battery	Island	Off-grid	No mention	COE
[23]	Diesel, PV, Wind, Hydro, Battery	Island	Grid	LF, CC	COE, NPC, RF
[24]	Diesel, PV, Battery	Island	Off-grid	LF, CC	COE, NPC, RF
[25]	PV, Battery	IIT Guwahati campus	Grid/Off-Grid	LF, CC	COE, NPC
[26]	PV, Wind, Biomass, Vanadium redox flow battery (VRFB)	IIEST	Off-Grid	No mention	COE
[28]	PV, Wind	Campus	Off-Grid	No Mention	NPC, COE, RF
[29]	PV, Fuel cell	Commercial Building	Off-Grid	LF, CC	COE
[30]	PV, Wind	High-rise building	Grid	No Mention	LCOE
[33]	PV, Wind	Residential Area	Off grid	CD	COE, NOC
[34]	Wind, PV, Biogas	City	Grid	No Mention	LCOE

LF: Load following; CC: Cycle charging; CD: Combined dispatch; PV: Photovoltaic; NPC: Net present Cost; COE: Cost of energy; RF: Renewable fraction.

presents a brief comparison of some of the recent research papers that focused on HRES design.

Several studies have declared renewable-based systems to be more economical and environmentally friendly than only-diesel or only-grid systems. In a grid-connected HRES study by Rad et al. [34], it was observed that the system cost of a stand-alone system was higher than the cost of a system with the same components and a grid connection. Therefore, renewable fraction and cost decrease by introducing grid connection in the same set of system components. Maximizing renewable penetration in grid-connected systems at the lowest possible cost is challenging, as most of the system’s energy requirements will be met by the grid if a connection is available. However, it is crucial to improve the renewable percentage in the production mix to curb environmental pollution.

It is observed that most of the hybrid system design studies were non-grid-connected and for remote rural locations. Furthermore, several studies can be found for cattle farms [35,36], agricultural lands [18,19], hilly regions [18], government organisations [37], microfinance institutions [11], office buildings [8], or islands [22,23,24,25,26,27]. Nevertheless, there is a lack of hybrid energy system design specifically for the oil and gas field area, supplementing the existing grid connection. Most of the concurrent studies are either designed using the default LF and CC dispatch strategies or there is no mention of the dispatch or design method at all. There is a lack of studies focused on design and development of the dispatch method customized for gas microturbine-, solar-, and river turbine-based grid connected systems. Accordingly, the present study focused on a grid-connected design, optimization, and modelling of HRES for oil and gas field areas utilizing the waste gas flares to produce power, thus minimizing environmental pollution from oil and gas production sites at the lowest possible cost. A novel dispatch strategy specifically designed to suit the requirements of oil and gas fields has been proposed. The presented research has the potential to convert oil and gas production sites from power-consuming and emission-causing sites to integrated energy companies. For the analysis, modelling, process and operation of a flare gas-based microgrid power management systems in the onshore oil and gas fields of Nazira-Assam, India, is discussed.

The rest of the paper is organized into following sections. Section 2 describes the mathematical models of all the components used in the HES design. This section forms the basic infrastructure of the study, as the HES installation locations and the previous studies on which the current study is based are briefly discussed in this section. Section 3 discusses the load profile of the area, and Section 4 presents the economic model, objective function, and environmental suitability indicators mathematically. In Section 5, the dispatch strategy customized to suit the case specific requirements is presented. Section 6 discusses various optimization constraints taken into account while designing the HES. Section 7 presents a technical, commercial, and ecological performance comparison of six different HESs designed for the two gas flare locations. Finally, the conclusion is given in Section 8.

2. Model Development and Design Specifications of HES Components

In [38], the authors selected two locations, namely Lakwa GGS5 and Geleky GGS2 to install HRES at the oil and gas field region of Nazira, Assam (27°1.8′ N, 94°52.4′ E), India. They concluded that Lakwa GGS5 was most suitable for installing any combination of HRES with a gas microturbine (GMT), whereas Geleky GGS2 was ideal for installing any combination of HRES with GMT and without a river turbine (RT). Based on those results, the current study proposed different combinations of retrofit hybrid energy systems (HES) for Lakwa GGS5 and Geleky GGS2. A 200 kW GMT is already installed at Lakwa GGS5 [39]. Therefore, as an extension to the previous studies [38,40], a techno–economic feasibility analysis for a retrofit system with the following general structure was performed for the two locations:

(1)

Geleky GGS2: GMT, Solar PV and Battery Energy Storage System (BESS);

(2)

Lakwa GGS5: GMT, Solar PV, RT, and BESS.

These components were selected as feasible preliminary technologies for the proposed case. In general, since the study area has abundant gas flare waste, as well as solar and water resources, a mixture of a gas microturbine, solar PV and a river turbine was considered for the HRES installations. Figure 1 mentions all the input parameters required to model various HES components, the optimization constraints and the overall configuration design plan.

2.1. Solar Photovoltaic

Solar PV panels were used to generate DC electricity, which was then fed into the AC system via power converters. To select the right PV panel, cost efficiency, tolerance to variation in power, conversion efficiency, and temperature coefficient are crucial factors. To estimate the performance of PV generators, IEC standard 61,724 and international energy agency PVPS task II guidelines were used [21]. Accordingly, CanadianSolar MaxPower CS6X-325P PV modules were chosen for the current study. As quoted by the seller and verified from the literature [41], the capital cost, replacement cost, and operation and maintenance cost of the PV panels were estimated to be 714 USD/kW, 700 USD/kW, and 14 USD/kW, respectively. The capital cost is more than the replacement cost, as it includes the cost of cables, inverters, installation, and substation charges. The selected PV modules had a derating factor of 88%, efficiency of 16.94%, and a nominal operating cell temperature (NOCT) of 45 °C [41]. The derating factor compensates for losses in the solar PV caused by dust, elevation, and drop in temperature, cable losses, etc., and makes the performance of rated PV close to real-life operation [42]. The efficiency of a PV panel decreases linearly with an increase in temperature [43]. The manufacturer’s rated performance was tested at 25 °C with NOCT of 45 °C, and the resultant 20 °C increase causes a reduction in rated power output [44]. Therefore, ambient temperature effects were considered while calculating the power output of the PV generation system using Equation (1) [45]:

$${\mathrm{P}}_{\mathrm{PV}}={\mathrm{Y}}_{\mathrm{PV}}\times \mathrm{d}\times \frac{{\mathrm{I}}_{\mathrm{T}}}{{\mathrm{I}}_{\mathrm{T},\mathrm{STC}}}\times \left[1+{\mathsf{\alpha }}_{\mathrm{p}}\left({\mathrm{T}}_{\mathrm{c}}-{ \mathrm{T}}_{\mathrm{c},\mathrm{STC}}\right)\right]$$

(1)

where ${\mathrm{Y}}_{\mathrm{PV}}$ is the rated capacity of the PV module (kW), $\mathrm{d}$ is the derating factor, ${\mathrm{I}}_{\mathrm{T}}$ is the incident solar radiations at the current conditions, ${\mathrm{I}}_{\mathrm{T},\mathrm{STC}}$ is the solar insolationon the PV panel at standard test conditions (STC) (1 kW/m²). ${\mathsf{\alpha }}_{\mathrm{p}}$is the temperature coefficient of power (%/°C), which accounts for the ambient temperature effects, and ${\mathrm{T}}_{\mathrm{c}}$ and ${\mathrm{T}}_{\mathrm{c},\mathrm{STC}}$ are the PV cell temperatures (°C) at real-time and at STC, respectively. The real-time PV cell temperature can be calculated using the following equation [42]:

$${\mathrm{T}}_{\mathrm{C}}={\mathrm{T}}_{\mathrm{a}}+\left({\mathrm{T}}_{\mathrm{C},\mathrm{NOCT}}-{\mathrm{T}}_{\mathrm{a},\mathrm{NOCT}}\right)\times \left(\frac{{\mathrm{I}}_{\mathrm{T}}}{{\mathrm{I}}_{\mathrm{T},\mathrm{STC}}}\right)\times \left[1-\left(\frac{{\mathsf{\eta }}_{\mathrm{mp}}}{\mathsf{\tau }\mathsf{\alpha }}\right)\right]$$

(2)

$${\mathsf{\eta }}_{\mathrm{mp}}={ \mathsf{\eta }}_{\mathrm{mp},\mathrm{STC}}\times \left[1+{ \mathsf{\alpha }}_{\mathrm{p}}\times \left(\frac{{\mathrm{T}}_{\mathrm{C}}}{{\mathrm{T}}_{\mathrm{C},\mathrm{STC}}}\right)\right]$$

(3)

where ${\mathrm{T}}_{\mathrm{a}}$ is the ambient air temperature (°C), ${\mathrm{T}}_{\mathrm{C},\mathrm{NOCT}}$ is the nominal operating cell temperature, ${\mathrm{T}}_{\mathrm{a},\mathrm{NOCT}}$ is the ambient temperature at which ${\mathrm{T}}_{\mathrm{C},\mathrm{NOCT}}$ is calculated (°C), ${\mathrm{T}}_{\mathrm{C},\mathrm{STC}}$ is the PV cell temperature at STC of 25 °C and ${\mathsf{\eta }}_{\mathrm{mp}}$ and ${\mathsf{\eta }}_{\mathrm{mp},\mathrm{STC}}$ are maximum power point efficiencies (in %) of the PV array at real-time and STC, respectively, $\mathsf{\alpha }$ is the coefficient of absorption of PV panel (%), and $\mathsf{\tau }$ is the transmittance associated with the PV panel cover (%).

Solar Irradiance

The global horizontal irradiance (GHI), which is the total daily solar radiation incident on a horizontal surface, for the chosen location (27°1.8′ N, 94°52.4′ E) was obtained from NASA surface meteorology and solar energy website [46]. The annual average solar radiation in the region is 3.92 kWh/m²/day. Figure 2a shows the average monthly values of the Clearness index (left axis) and GHI (right axis) for 22 years (July 1983–June 2005). The clearness index (surface radiation divided by the extraterrestrial radiation) is the measure of the clarity of the atmosphere and is a dimensionless quantity varying between 0 and 1. The clearness index is high in winter, whereas GHI is higher in the summer months. Both these indices are used to calculate the flat-plate PV output.

2.2. River Turbine

Hydrokinetic River Turbines (HKRT) require a smaller amount of civil engineering and construction works as compared to hydel power stations and dams [47]. The planned installation locations (Lakwa GGS5 and Geleky GGS2) are swept by rivers Dikhow and Disang, which are tributaries of the river Brahmaputra and have average river heads varying from 2 to 5 m [48]. The most common problem with such low river heads is the rotor blockage due to debris deposition; axial flow turbines with blades that swing back and shed off the debris are considered to be most effective in dealing with this problem [49]. Deployment of an HKRT becomes difficult in deep and high flow velocity rivers, and darreius turbines are a good fit for such cases [49]. Since the river head in the present case is low (2 to 5 m), and flow velocity is is as low as 3.152 m/s in the summers and as high as 19.8 m/s in the rainy season (Figure 2b), the axial flow Schottel hydrokinetic turbine manufactured by GmbH Germany is the most suitable fit [49,50]. It has a rotor diameter of 3 m, rated power of 70 kW, rated water velocity of 3.8 m/s, cut-in speed of 0.9 m/s, and a cut-out speed of 6.75 m/s [51]. The lifetime of the selected turbine is 25 years [51]. The hydrokinetic turbine power curve in Figure 3a derived from the manufacturer’s datasheet [50] shows that the maximum output is obtained if the water speed is between 3.75 and 6.75 m/s, and a water speed above 0.89 m/s will start giving exponentially increasing power output. Figure 3b shows the installation of a schottel hydrokinetic turbine.

The hydro turbine’s nominal power output is given by Equation (4) [42].

$${\mathrm{P}}_{\mathrm{hyd}}={ \mathsf{\eta }}_{\mathrm{hyd}}{\mathsf{\rho }}_{\mathrm{water}}{\mathrm{gh}}_{\mathrm{net}}{\mathrm{Q}}_{\mathrm{turbine}}$$

(4)

where ${\mathsf{\eta }}_{\mathrm{hyd}}$ is the turbine efficiency, ${\mathsf{\rho }}_{\mathrm{water}}$ is the density of water, g is the gravitational acceleration, ${\mathrm{h}}_{\mathrm{net}}$ is the net head, and ${\mathrm{Q}}_{\mathrm{turbine}}$ is the flow rate through the turbine [51] expressed as,

$${\mathrm{Q}}_{\mathrm{turbine}}=\frac{1}{2}.\frac{{\mathrm{C}}_{\mathrm{T}}. \mathrm{A}.{ \mathrm{v}}^3}{\mathrm{g}.{ \mathrm{h}}_{\mathrm{net}}}$$

(5)

where ${\mathrm{C}}_{\mathrm{T}}$ is the coefficient of power of the hydrokinetic turbine, $\mathrm{A}$ is the swept by turbine blades, and $\mathrm{v}$ is the velocity of river water in m/s [47]. Therefore,

$${\mathrm{P}}_{\mathrm{hyd}}=\frac{1}{2}.{\mathsf{\eta }}_{\mathrm{hyd}}.{ \mathsf{\rho }}_{\mathrm{water}}. {\mathrm{C}}_{\mathrm{T}}. \mathrm{A}.{ \mathrm{v}}^3$$

(6)

2.3. Converter

The converter works as a coupling between DC and AC systems and can facilitate the bidirectional power flow. Since the proposed model has a solar PV and river turbine modeled on the DC bus, and a gas microturbine, grid, and load modeled on the AC bus, a bidirectional converter with an input efficiency of 95% and rectifier input efficiency of 97% was considered [40].

2.4. Gas Microturbine

The average annual flaring of natural gas in the area under consideration is 2958.25 MMSCM (million metric standard cubic meters) daily [40]. The gas flaring amounts to 11.833 million tons of CO₂ emissions and a loss of USD 59,165 annually [3]. The aim of this study was to utilize the waste flare gas to produce electricity, thereby reducing carbon emissions and the electricity problem in an economically optimal way. Therefore, three Capstone gas microturbines were employed for the hybrid renewable energy system design: two 30 kW gas MTs with a fuel curve intercept of 1.50 m³/h and a 200 kW gas MT with a fuel curve intercept of 10.0 m³/h. Both types of gas MTs have a fuel curve slope of 0.29 m³/h/kW (Figure 4) [40].

The fuel curve, as shown in Figure 4, is a straight line with a y-intercept, and it uses the following equation for the fuel consumption of the generator:

$$F=F_0{\mathrm{Y}}_{\mathrm{gen}}+F_1{\mathrm{P}}_{\mathrm{gen}}$$

(7)

where $F_0$ is the fuel curve intercept coefficient, $F_1$ is the fuel curve slope, and ${\mathrm{Y}}_{\mathrm{gen}}$ is the rated capacity of the generator (in kW).

The fixed and marginal energy costs of the generator estimated by Equations (8) and (9), respectively, were used for simulation of the system operation. The fixed cost is simply the cost of running a generator without generating electricity, whereas the marginal cost is the additional cost per kilowatt-hour of electricity produced by the generator:

$${\mathrm{C}}_{\mathrm{gen},\mathrm{fixed}}={ \mathrm{C}}_{\mathrm{om},\mathrm{gen}}+\frac{{\mathrm{C}}_{\mathrm{rep},\mathrm{gen}}}{{\mathrm{R}}_{\mathrm{gen}}}+F_0{\mathrm{Y}}_{\mathrm{gen}}{\mathrm{C}}_{\mathrm{fuel},\mathrm{eff}}$$

(8)

where ${\mathrm{C}}_{\mathrm{om},\mathrm{gen}}$ is the operation and maintenance cost (in USD per hour), ${\mathrm{C}}_{\mathrm{rep},\mathrm{gen}}$ is the replacement cost in USD, ${\mathrm{R}}_{\mathrm{gen}}$ the lifetime of the generator (in hours), $F_0$ is the fuel curve intercept coefficient or the quantity of fuel per hour per kilowatt, ${\mathrm{Y}}_{\mathrm{gen}}$ is the capacity of the generator (in kW), ${\mathrm{C}}_{\mathrm{fuel},\mathrm{eff}}$ is the valid fuel price (in USD per quantity of fuel).

$${\mathrm{C}}_{\mathrm{gen},\mathrm{mar}}={ F}_1{\mathrm{C}}_{\mathrm{fuel},\mathrm{eff}}$$

(9)

where $F_1$ is the fuel curve slope in the quantity of fuel/h/kW, and ${\mathrm{C}}_{\mathrm{fuel},\mathrm{eff}}$ is the effective price of fuel in USD/quantity of the fuel.

2.5. Grid

The grid was modeled as a component that allows micro power systems to purchase and sell AC electricity in case of a power shortage or surplus production. Whether the power will be sold to the utility or purchased from it is decided using the dispatch strategy discussed in Section 5. The rate of selling back power to the grid was set as 0.108 USD/kWh, and the power purchase price was set to 0.081 USD/kWh [52]. The capacity ratings and cost specifications of various components used are summarized in

Table 2. Various costs, lifetime, type, and rating details of various components involved in the HES design [40].

Component Name	Type	Rating	Quantity	Capital Cost	Replacement Cost	O&M Cost	Lifetime
Gas Microturbine	Capstone	200 kW	1	USD 200,000	USD 200,000	0.005 USD/operational h	40,000 h
		30 kW	2	USD 30,000	USD 30,000	0.005 USD/operational h	40,000 h
Solar Photovoltaics	CanadianSolar MaxPower	2000 kW	PV arrays arranged in series and parallel	0.08 USD/kWh	0.08 USD/kWh	-	25 years
Hydrokinetic River Turbine	Schottel	70 kW	1	USD 57,000,000	USD 57,000,000	USD 6,290,000	25 years
Converter	Free size	10,455 kW 1	1	USD 375	USD 375	USD 10	25 years

¹ Fixed after optimization.

.

3. Load Demand Estimation

The Assam Power Distribution Company Limited (APDCL) supplies power to residential customers, commercial customers, and industries in the study area. Hence, the site has a mixed hourly load profile as shown in Figure 5. The average annualized load demand of the Nazira oil and gas field region is 16,500 kWh/day, with a peak load of 2040.49 kW for the year 2019 as obtained from APDCL. The regional power grid operates on a power factor of 0.85 and a load factor of 0.34. To make the load data more accurate, typical daily load profiles for all months were specified, and then the randomness was added to synthesize the data. A daily variability of 10% was selected at every 20% variation in time-step [53]. It can be seen from Figure 5 that loading occurs mostly between 7:00 a.m. and around 11:00 p.m., and peak load occurs between 6:00 p.m. and 11:00 p.m. The remaining hours of the day experience light loads. The grid-connected Microgrid power management system has to be scheduled accordingly.

4. Economic Modelling and Objective Function

The net present cost (NPC) incurred during the project lifetime is the best indicator of the suitability of a given HRES design. The NPC calculation takes all possible factors into account: interest rate, inflation, discount rate, capital costs, operation and maintenance costs, replacement, and salvation costs. The objective function minimizes the NPC, the expression for which is given by Equation (10) [54]:

$$Obj Fn=\min \left(\mathrm{NPC}\right)=\min \left(\frac{{\mathrm{C}}_{\mathrm{tot},\mathrm{ann}}}{\mathrm{CRF} \left(d, \tau \right)}\right)$$

(10)

where ${\mathrm{C}}_{\mathrm{tot},\mathrm{ann}}$ is the total annualized cost of the system (USD/year) and CRF is the capital recovery factor, which depends on discount rate ($d$) and project lifetime ($\tau$, years) as expressed by Equations (11) and (12):

$$\mathrm{CRF} \left(d, \tau \right)=d\times \frac{{\left(1+d\right)}^{\tau }}{\left[{\left(1+d\right)}^{\tau }-1\right]}$$

(11)

where $d$ is expressed as,

$$d=\frac{\left(d^{\prime }-f\right)}{\left(1+f\right)}$$

(12)

where $d^{\prime }$ is the nominal discount rate, and $f$ is the expected inflation rate.

The capital cost of each component is converted to the annual cost by amortizing it over the component’s life using the actual discount rate. The total annualized cost of the system was estimated by Equation (13):

$${\mathrm{C}}_{\mathrm{tot},\mathrm{ann}}={\displaystyle \sum }_{\mathrm{k}\in \mathrm{PV}, \mathrm{GMT}1, \mathrm{GMT}2,\mathrm{RT}, \mathrm{Converter} }{\mathrm{C}}_{\mathrm{cap},\mathrm{k}}+{\mathrm{C}}_{\mathrm{O}\&\mathrm{M},\mathrm{k}}+{\mathrm{C}}_{\mathrm{repl},\mathrm{k}}-{\mathrm{C}}_{\mathrm{salv},\mathrm{k}}$$

(13)

where Cs_alv is the salvage cost, C_cap is the capital cost, C_O&M is the operation and maintenance cost, C_repl is the replacement cost, and k denotes the system components—PV, GMT, RT, and converter.

The salvage cost of a system component is directly proportional to its remaining life and is dependent on the component’s replacement cost rather than capital investment cost. The salvage cost is given by Equation (14) [40].

$${\mathrm{C}}_{\mathrm{salv},\mathrm{k}}={\mathrm{C}}_{\mathrm{repl},\mathrm{k}}\left(\frac{{\mathrm{T}}_{\mathrm{rem},\mathrm{k}}}{{\mathrm{T}}_{\mathrm{com},\mathrm{k}}}\right)$$

(14)

Here, ${\mathrm{T}}_{\mathrm{rem}}$ is the remaining life of the component in years, and ${\mathrm{T}}_{\mathrm{com}}$ is the component lifetime in years. The system with a minimum life-cycle cost ultimately results in a system with a minimum Levelized cost of energy (LCOE). LCOE is the average cost per kWh of the useful electricity generated by the system and is given by Equation (15) [40].

$$\mathrm{LCOE}=\frac{{\mathrm{C}}_{\mathrm{tot},\mathrm{ann}}}{{\mathrm{E}}_{\mathrm{tot},\mathrm{ann}}}$$

(15)

${\mathrm{E}}_{\mathrm{tot},\mathrm{ann}}$ is the total amount of energy used in meeting the load demands per year (kWh/year). Additionally, the return on investment (ROI) is another crucial indicator, which determines the profitability or economic viability of the HES as an investment as expressed in Equation (16) [54]:

$$\mathrm{ROI}=\frac{{\sum }_{\mathrm{i}=1}^{\mathsf{\tau }}\left({\mathrm{C}}_{\mathrm{i},\mathrm{ref}}-{ \mathrm{C}}_{\mathrm{i},\mathrm{sel}}\right)}{\mathsf{\tau }\times \left({\mathrm{C}}_{\mathrm{cap},\mathrm{sel}}-{ \mathrm{C}}_{\mathrm{cap}, \mathrm{ref}}\right)}$$

(16)

where, ${\mathrm{C}}_{\mathrm{i},\mathrm{ref}}$ and ${\mathrm{C}}_{\mathrm{i},\mathrm{sel}}$ are the nominal annual cash flow of the base/reference and selected systems, respectively, and ${\mathrm{C}}_{\mathrm{cap},\mathrm{ref}}{ \mathrm{and} \mathrm{C}}_{\mathrm{cap},\mathrm{sel}}$ are the corresponding capital costs.

Environmental Suitability Indicators

The renewable fraction (RF) and amount of CO₂ emitted (A_CO2) were chosen as the sustainability indicators as expressed by Equations (17) [55] and (18) [54], respectively:

$$\mathrm{RF}=1-\frac{\mathrm{Enon}\_\mathrm{ren}}{\mathrm{Etot},\mathrm{ann}}$$

(17)

where $\mathrm{Enon}\_\mathrm{ren}$ is the non-renewable generation in kWh/yr, and $\mathrm{Etot},\mathrm{ann}$ is the total electrical load served (in kWh/yr), which included the gas MT and Grid in the present case.

$${\mathrm{A}}_{\mathrm{CO}2}=3.667\times {\mathrm{m}}_{\mathrm{f}}\times {\mathrm{LHV}}_{\mathrm{f}}\times {\mathrm{CEF}}_{\mathrm{f}}\times {\mathrm{X}}_{\mathrm{c}}$$

(18)

where ${\mathrm{m}}_{\mathrm{f}}$ is the amount of diesel fuel (L), ${\mathrm{LHV}}_{\mathrm{f}}$ is the fuel heating value (MJ/L), ${\mathrm{CEF}}_{\mathrm{f}}$ is the factor related to carbon emission (ton carbon/TJ), and ${\mathrm{X}}_{\mathrm{c}}$ is the oxidized carbon fraction, where each 3.667 gm of CO₂ includes 1 gm of carbon.

5. Dispatch Strategy

A dispatch strategy is a logical sequence of instructions that govern the operation and flow of energy between various HRES components. After accounting for the current power/energy status of the component, it determines which equipment must operate at that instant. HOMER Pro software uses cycle-charging (CC) and load-following (LF) strategies by default [42]. In both strategies, renewable-source-based generators are given top priority while meeting the load. The basic difference between CC and LF is that, in CC, the generator is used to meet the load as well as charge the battery, while it is used only to meet the load in LF. The battery is charged using surplus renewables in the case of LF depending on the renewable availability [56]. Since the use of the generator is less in LF, it is more effective in minimizing the environmental pollution and overall system operating cost (as the cost of fuel will be reduced) [57]. However, in the current case, as the natural-gas-based generator uses flare gas waste to generate electricity, the gas-microturbines are considered equivalent to renewable-source-based generators. Therefore, the dispatch strategy was customized to suit the requirements of the presented case study. Equations (19) and (20) form the starting points of the dispatch strategy calculations.

$${\mathrm{P}}_{\mathrm{Ren}}={ \mathrm{P}}_{\mathrm{PV}}+{\mathrm{P}}_{\mathrm{RT}}+{\mathrm{P}}_{\mathrm{MT}}$$

(19)

$${\mathrm{P}}_{\mathrm{X}}={ \mathrm{P}}_{\mathrm{L}}-{\mathrm{P}}_{\mathrm{Ren}}$$

(20)

where ${\mathrm{P}}_{\mathrm{Ren}}$ is the total renewable generation, which is the sum of the power produced by solar PV (${\mathrm{P}}_{\mathrm{PV}}$), hydrokinetic river turbine (${\mathrm{P}}_{\mathrm{RT}}$), and gas microturbine (${\mathrm{P}}_{\mathrm{MT}}$); ${\mathrm{P}}_{\mathrm{X}}$ is the outstanding power requirement after utilizing ${\mathrm{P}}_{\mathrm{Ren}}$ to meet the load demand (${\mathrm{P}}_{\mathrm{L}}$). Thus, decisions must be made depending on the value of ${\mathrm{P}}_{\mathrm{X}}$, as illustrated in Figure 6. ${\mathrm{P}}_{\mathrm{X}}=0$ indicates ${\mathrm{P}}_{\mathrm{L}}$ was fully met by renewables, while ${\mathrm{P}}_{\mathrm{X}}>0$ means there is outstanding power requirement, which has to be met from the grid. However, ${\mathrm{P}}_{\mathrm{X}}<0$ indicates the availability of surplus renewable power, which can be fed to the grid.

The algorithm shown in Figure 6 runs for the entire year and makes hourly calculations for real-time dynamic power dispatch between various components (8760 h in a year). Whenever the system is deficient in power, it can fetch power from local generation and the grid using the proposed dispatch strategy, making the system robust and reliable.

6. Optimization Constraints

The following constraints and assumptions were followed in the present case study:

(1)

Resource constraints: the selected study location has annual solar radiation of 3.14 kWh/m²/day [46], and annual average river flow of 3.152 m/s and 19.8 m/s in summers and winters, respectively.

(2)

Economic: the nominal discount rate determines the present lump sum value of future cash flows. A nominal discount rate of 8% was considered for the present study. Considering the scale of the current operational project in ONGC Nazira [39] and the availability of the natural gas reserves in the area [58], a project lifetime of 25 years was considered.

(3)

Technical: an operating reserve of 10% was selected as a percentage of the load. This accounts for abrupt changes in load patterns, and an active operating reserve of 50% was considered for solar photovoltaic to deal with the intermittent weather conditions.

(4)

Reliability: to ensure high levels of power supply reliability, the ‘caidi’ index—the average number of interruptions per year—was considered 0%.

(5)

Emission: currently, there is no law on emission penalties in Assam. Therefore, the emission penalties for CO₂, nitrogen oxides, carbon monoxide, particulate matter, sulphur dioxides, were assumed to be zero.

7. Results and Discussion

Six configurations of HES were selected for two different locations: Lakwa and Geleky. Since Geleky GGS2 is away from the river stream, configurations devoid of hydrokinetic river turbines were selected for Geleky GGS2, while configurations with a river turbine were suggested for Lakwa GGS4 owing to its vicinity to the river stream. The six configurations were compared on the basis of (a) cost: NPC, LCOE, operating expenses, and (b) environmental suitability: total fuel consumption, renewable fraction, as shown in

Table 3. Comparison of the selected HES based on cost and environmental sustainability.

#	Configuration	Cost Parameters			Environmental Sustainability
		NPC	LCOE	Operating Cost	Renewable Fraction (%)	Total Fuel Consumption (m3/yr)	Suggested Location
1	PV/GMT/RT/G	USD −2.32B	USD −2.72	USD −187M	93.8	774,384	Lakwa
2	GMT/RT/G	USD 571B	USD 7.16	USD 39.8M	1.99	774,384	Lakwa
3	PV/RT/G	USD −2.14B	USD −2.56	USD −173M	95.6	0	Lakwa
4	PV/GMT/G	USD −2.45B	USD −2.87	USD −192M	93.7	774,384	Geleky
5	GMT/G	USD 445M	USD 5.61	USD 34.4M	0	774,384	Geleky
6	PV/G	USD −2.27B	USD −2.72	USD −178M	95.5	0	Geleky

.

System numbers 2 and 5—GMT/RT/G and GMT/G have exorbitant net present, operating costs as well as levelized cost of energy (LCOE) (NPC USD 571B and USD 445M, operating cost: USD 39.8M and USD 34.4M, LCOE: USD 7.16 and USD 5.61). It can be observed from Table 3 that all systems except GMT/RT/G and GMT/G (G denotes grid) are profitable as they have negative cash flows. Lower cost values (less positive and more negative) imply more grid exports [59]. Furthermore, the renewable fraction available in these two systems is negligible (1.99% and 0%). Therefore, GMT/RT/G and GMT/G were excluded from the viable options, and a cost comparison of the remaining four systems was made, as shown in Figure 7. The NPC represents the overall lifecycle cost of the system with all the future cashflows discounted back to the present [40]. Out of all four systems, PV/GMT/G has the lowest NPC. It may be noted from Table 3 that all four systems have almost equally renewable fractions. Therefore, PV/GMT/G is the best suitable system economically and environmentally for the considered oil and gas field location.

Furthermore, an analysis of the component-wise electrical power consumption and emissions from the operation of the six HES for a lifetime of 25 years was performed, and the results are shown in Figure 8, Figure 9 and Figure 10, respectively. Figure 8 shows that the presence of RT does not have any significant impact on the share of emissions in any HES. It also depicts that the proportion of CO₂, CO, and nitrogen oxides is almost constant for PV/G–PV/RT/G, GMT/G–GMT/RT/G, and PV/GMT/G–PV/GMT/RT/G.

It can also be observed from Figure 8 that the electrical power production from RT in the systems with RT is negligible, whereas the distribution of power produced by GMT, PV, RT, and G is almost constant for PV/G–PV/RT/G, GMT/G–GMT/RT/G, and PV/GMT/G–PV/GMT/RT/G systems. A detailed analysis of emissions from the six systems is provided in Figure 9. It is observed that both GMT, G and GMT, RT, G cause the highest amount of overall emissions, making them unfit for the environment. With respect to overall emissions, PV, GMT, G, and PV, GMT, RT, G produced the lowest values. However, by considering the nominal contribution of RT in the total electrical power production, PV, GMT, G could be finalized as the most viable configuration for both the locations.

7.1. Production Analysis of the Most Optimal Solution (PV/GMT/G)

A combination of solar PV, two 30 kW and one 200 kW GMTs, and a grid was observed to be the most optimal configuration. Figure 11 provides the monthly component-wise electrical power production summary of the PV/GMT/G system. The maximum power is generated in August, followed by July, May, June, and January. Likewise, the maximum power exchange with the grid also occurs in mid-year (i.e., June to August). The electrical penetration by PV in the entire system comes out as 52.3%. Figure 12 depicts that the load consumes 79.8% of energy, whereas 20.2% is sold off to the utility power grid. Figure 13 provides a detailed image of the energy exchange with the grid by the winning system architecture. It can be observed that the load demand is comparatively less in winter months (Nov, Dec, Jan, Feb) than in the other months. Therefore, surplus energy is sold to the grid during the winter months. However, the peak demand is the highest (2036 kW and 2093 kW) in the summer months (June–July). It can also be observed that the energy cost is at its maximum in August. Figure 14 shows that power has to be purchased from the grid to meet the surplus load mainly during night hours (5:00 p.m. to 11:00 p.m.), and surplus energy is sold back to the grid during the daytime (6:00 a.m. to 5:00 p.m.).

7.2. Comparison of the Proposed System with the Existing Power System

Compared to the existing power system, the proposed GMT/PV/G design is 35.54% economically (USD 5.9M reduction in NPC) and 50% ecologically (decrease of 850 tonnes of CO₂ per year) more sustainable. Figure 15 shows a comparison of the cumulative as well as annualized cash flows by the two systems. The downward slope of the cumulative nominal cash flow graph indicates expenditures. A smaller negative slope implies that setting up a hybrid system in the Nazira area is more profitable than the existing grid power supply, calculated for 25 years.

After a payback period of 7.8 years, the designed hybrid system proves out to be profitable with an ROI (return on investment) of 11%. The lifetime for GMT is 40,000 h, and hence there occurs a replacement cost for the gas microturbines after completing 40,000 h, so there will be a replacement cost for GMT after this period (5th year). Therefore, there would be a spike in the annual nominal cash flow for both the proposed and existing systems in the fifth year. The existing system’s yearly cash flow becomes higher as INR 40 million per year must be spent for operation and maintenance throughout the lifetime of the project (25 years). Consequently, the overall cost of the existing system supersedes that of the proposed hybrid system.

Major emissions from the natural gas generators and the designed system include CO₂, CO, unburnt hydrocarbons, particulate matter, sulfur dioxide, and nitrogen oxides. Figure 16 provides an exact statistical comparison of emissions from the hybrid system with the existing system. Compared to the existing system, the emissions of CO₂, CO, particulate matter, sulfur dioxide, and nitrogen oxides were reduced by 24.52%, 13.05%, 12.90%, 41.21%, and 8.52%, respectively. The CO₂ emissions from the proposed hybrid system is up to 50% less than that from direct flaring (statistics of the CO₂ emissions from direct flaring provided in [60]). Therefore, this study proves that the combination of gas micro turbines with PV, to generate electricity from natural gas, reduces emissions by 40–50%.

8. Conclusions

Due to gas flaring at oil and gas fields, the flora–fauna and human health of settlements around the flaring site are adversely affected. Another issue is the lack of a reliable and robust power supply in the oil and gas field area. This research work proposed a hybrid energy system (HES) that utilizes waste gas flares to generate electricity, thereby solving both the aforementioned problems. A systematic framework was established to suggest the optimal size and design of the HES. A novel dispatch algorithm was used to simulate the real-time operation of the HES to suit the requirements of the oil and gas fields. The techno–commercial and environmental viability of six HESs was analyzed for two different gas flaring sites. A system comprising two 30 kW gas microturbines, 2000 kW solar PV panel set, and grid connection could save 35.52% net present cost and reduce gas flaring by up to 50%. This research is in line with the United Nations Sustainable Development Goals and can be beneficial for the engineers and researchers who aim to improve productivity while reducing gas flaring in oil and gas fields.