Current World Environment

Introduction

Diesel-powered vehicles have been gaining significant market share in the Indian automotive industry, owing to their inherent advantages, including enhanced efficiency and power. However, the simultaneous reduction of NOx (Nitrogen oxides) and PM/soot emissions while maintaining optimal engine performance and adhering to rigorous pollution regulations presents a formidable challenge for the automobile sector ¹. Many researchers have focused towards the utilization of ethanol as an alternative for diesel fuel. This is because of ethanol’s renewability, biodegradability, and its capacity to reduce greenhouse gas emissions, making it a promising alternative in the pursuit of environmentally sustainable fuel options ^2–4. To utilise ethanol in CI engines, researchers proposed many techniques, including diesel-ethanol blending, diesel-ethanol-additives blending, ethanol-diesel fumigation. The diesel-ethanol blending technique has been found to improve engine efficiency while reducing soot and PM emissions. However, it can lead to increased emissions of NOx, HC, and CO ^5–7. On the other hand, increasing the proportion of diesel-to-ethanol replacement may lead to a more significant simultaneous reduction in PM/soot and NOx in CI engines ⁸. However, this enhancement in emissions control might come at the expense of engine performance, as noted in previous studies ^5,6,9. The ethanol fumigation approach has demonstrated a reduction in NOx, soot, and PM emissions, but it has the opposite effect of increasing HC and CO emissions, which is in contrast to the impact of diesel-ethanol blending. Furthermore, this approach has been associated with a greater pressure rise rate and an extended ignition delay period, both of which can lead to decreased engine performance, including reduced brake thermal efficiency and uncontrolled combustion ^10–17. 

Consequently, to maximize the benefits of both blending and fumigation, researchers have proposed the simultaneous use of ethanol in two fuel induction techniques (combined blending and fumigation) ¹¹. However, it's important to note that the results obtained through this combined approach do not reach the same extent as those observed in either the blending or fumigation technique alone. The incorporation of metal-based nanoparticles into diesel-ethanol or diesel-ethanol-biodiesel blends has shown promising results. This addition increases energy efficiency, combustion efficiency, and heat release rates due to the catalytic action of the nanoparticles. Consequently, this leads to a reduction in NOx, HC, and CO emissions. Moreover, it significantly improves engine performance by reducing ignition delay periods ^4,18–24. Additionally, the incorporation of metal-based nanoparticles into these fuel blends enables the utilization of a higher fraction of ethanol, resulting in a more substantial reduction in PM or soot emissions. ^25,26.

Consequently, there is a need for experimental investigations to be conducted to assess the impact of ethanol fuel induction techniques, both with and without the use of metal oxide-based nanoparticles like ZnO. These experiments should focus on evaluating emissions, engine performance, combustion characteristics, and exhaust after-treatment economy under different operating loads and conditions of CI engines. 

The diverse outcomes observed in these experiments pose a challenge in identifying the optimal ethanol fuel techniques and fuel blends (diesel-ethanol or diesel-ethanol-nanoparticles) across a range of engine loads, including low, medium, and high. The evaluation of these fuel induction techniques and fuel blends involves a comprehensive assessment across multiple criteria, including performance, combustion characteristics, emissions, and after-treatment economy. Moreover, these evaluations are conducted under different operating conditions, spanning low, medium, and high engine loads. There are studies that utilize the multi-criteria decision-making (MCDM) method to develop comprehensive model to identify the most suitable alternative fuels under different operating conditions. MCDM is a fundamental approach to addressing decision-making challenges, particularly when the goal is to identify the optimal alternative by taking into account multiple criteria in the selection process ^27,28.  In the study conducted by Sakthivel et al., a hybrid MCDM technique is implemented to optimize the selection of the biodiesel blend (Fish oil) from a set of six alternative fuel blends. The alternatives encompass pure diesel, B20, B40, B60, B80, and B100. This hybrid MCDM approach integrates the analytic network process (ANP) for assigning weights to evaluation criteria through pairwise comparisons. Subsequently, the technique for order of preference by similarity to ideal solution (TOPSIS) is employed to analyze the data and rank the alternative fuel blends²⁹. Sinan Erdo?an et.al. studied the application of innovative hybrid MCDM techniques, specifically step wise weight assessment ratio analysis (SWARA) - multi-objective optimization on basis of ratio analysis (MULTIMOORA) and ANP - MULTIMOORA, to determine the best fuel for CI engines. Through this analysis, it was concluded that VOB20 (vegetable oil biodiesel with a 20% blend) emerged as the top-performing fuel, while AFB100 (animal fat biodiesel with a 100% blend) ranked as the least favourable option among the nine types of fuels considered. The fuels considered in the study comprised diesel, biodiesel derived from vegetable oil, biodiesel derived from animal fat, and various blends of these fuels ²⁷. Sinan Erdo?an et.al. studied on the application of the MCDM technique model known as operational competitiveness rating analysis (OCRA). This technique is used to assess and evaluate the test results by comparing data related to performance, combustion, and emission criteria between a thermal barriers coated (TBC) diesel engine and an uncoated standard (STD) engine. In the OCRA method, the weights allocated to the criteria can be determined through analytic hierarchy process, SWARA, or a simpler method based on subjective assignment, where experience plays a significant role in determining the weights ³⁰. M.K. Balki et al. employed a hybrid method known as SWARA-ARAS (additive ratio assessment) to determine the optimal operating parameters. These parameters included three distinct factors: ignition timing, compression ratio, and air-fuel ratio. The study entailed experimental investigations conducted across 81 varying conditions. The results of this research facilitated the identification of the most optimal operating parameters for the utilization of methanol fuel, which include a compression ratio of 9:1, an air-fuel ratio of 1.1, and an ignition timing of 20⁰ CA³¹.

The literature reveals that MCDM methods are a valuable approach for investigating and optimizing working conditions based on multiple criteria when using alternative fuels in engines. Several studies have employed these methods, often in hybrid combinations, to evaluate results comprehensively. These studies have used various methods, including ANP, AHP, and SWARA, to determine criteria weights. Optimization techniques like TOPSIS, PROMETHEE, and MULTIMOORA have then been applied to make informed decisions. However, the use of hybrid models in MCDM is relatively limited. Specifically, there is a lack of literature regarding the use of a hybrid model combining analytic hierarchy process-AHP and preference ranking organization method for enrichment evaluations-PROMETHEE II for selecting optimal operating conditions or the best alternative fuel blends for internal combustion engines. 

The aim of this study is to establish a decision method that assesses various ethanol fuel induction techniques and blends across a range of operating loads. Employing a hybrid approach, this study combining AHP for criteria weighting and PROMETHEE II to rank the alternatives. The practical application of this approach is directed towards the development of ethanol fuel induction control strategies. These strategies are designed to enhance the cleanliness, performance, and cost-effectiveness of compression ignition (CI) engine operation across a spectrum of load conditions, including low, medium, and high engine loads. The study aims to contribute to advancements in CI engine performance by offering insights into the most effective ethanol fuel configurations and induction methods for various operational demands.

Materials and experimental methodology

Test fuels and experimental setup

The fuels utilised in this study were commercially available diesel and ethanol (purity> 99.5%). In addition, ZnO- nanoparticles (100 ppm, 20nm) were used as an additive in a diesel-ethanol (DE10) blend to prepare diesel-ethanol-nanoparticle (DE10_NP) blend. The properties of diesel, ethanol, and DE10 were measured and reported in Table 1 using ASTM standards for these investigations. In this study, single-cylinder CI engine was tested using several fuelling techniques, such as blending (DE10), ethanol fumigation (Fumigation), and combined blending-fumigation (DE10 + Fumigation), with and without the use of nanoparticles. 

Table 1:  Properties of test fuels

The experimental configuration comprises a single-cylinder, four-stroke CI research engine with an eddy current dynos, for loading. The engine's specifications are presented in Table 2. The research engine included an additional fuel supply system with ethanol injection to the intake manifold for ethanol fumigation. Exhaust gases were also sampled in the tailpipe for HC, CO, NOx, and smoke emission analyses. The HC, CO, and NO_x were measured with an AVL DIGAS 444N 5-gas analyzer and the smoke value with an AVL 437 smoke meter. Table 3 lists the specifications of the exhaust gas analyzer and smoke meter. Fig.1 displays a diagram of the CI engine experimental arrangement, the ethanol fumigation system, and exhaust measurements.

Table 2: Engine specifications

Table 3: Specifications of gas analyzer and smoke meter

Figure 1: Schematic diagram of CI engine test rig along with ethanol fumigation system   Click here to view Figure

Experimental procedure for data collection 

Several test runs were conducted on a single cylinder compression ignition engine under various operating conditions. Observations of engine performance, combustion, and emission attributes were conducted at low (0-20%), medium (40-60%), and high (80-110%) engine loads using blending (DE10), ethanol fumigation (Fumigation), and combined blending-fumigation (DE10 + Fumigation) fuel testing techniques with DE10  and DE10_NP fuel blends. After several engine runs at each engine load, the desired quantity of ethanol (10% in blending and 30% in fumigation, combined blending-fumigation) and diesel at each engine load was pre-calibrated by adjusting the diesel pump sleeve and ethanol injection was controlled by the ECU based on input from external sensors and the testing parameters.

MCDM methodology

The objective of this study is to select the optimal ethanol fuel induction technique for various engine loads based on diverse criteria, employing a two-step approach that combines two distinct multi-criteria decision-making methods. The proposed hybrid model, consisting of AHP-PROMETHEE II, operates in two phases: firstly, the AHP method determines the criteria weights, and secondly, the PROMETHEE II method ranks alternatives in descending order of preference. The methodology proposed is elaborated in the subsequent sections.  

Selection of alternatives 

In this experiment, the engine was operated across a range of loads, including low load (0-20%), medium load (40-60%), and high engine load (80-110%), using various fuels (diesel, DE10, DE10_NP) and different fuelling techniques (blending, fumigation, and combined blending-fumigation). At each load range, combustion performance, emissions, and after-treatment requirements are expected to vary depending on the choice of fuels and fuelling techniques. Experimental studies comprised a total of 18 distinct alternatives, each coded as per the definitions outlined in Table 4.

Table 4: No. of alternatives and their coding

Criteria identification

This section involves identification of criteria to evaluate 18 alternative for optimization of ethanol fuel induction techniques and fuel blends. This comprehensive assessment considered 18 specific criteria would be taken into account, including maximum cylinder pressure (CPMAX), maximum rate of pressure rise (RPR MAX), net heat release (NHR), combustion duration (CD), delay period (DP), NOx, SOOT, HC, CO, CO2, exhaust temperature (EGT), brake thermal efficiency (BTE), brake specific fuel consumption (BSFC),brake power (BP), volumetric efficiency (VOLE), urea consumption for SCR (UC), post fuel consumption for regeneration (PFC), and relative engine operating cost (REOC). These criteria values against each alternatives were calculated and measured from the engine test results to form initial data value matrix as shown in Table 5. Criteria such as SCR-out NOx, high aqueous urea consumption, fuel or energy requirements for on-board active regeneration of the particle filter, and relative operating expenses of engine after-treatment systems were calculated using Eq. (1-6) to estimate after-treatment economy, with the corresponding data values provided in Table 5. 

Table 5: Data of the criteria according to alternatives   Click here to view Table

AHP decision model

AHP is a methodical decision technique designed to assist individuals or groups to deal with complex decisions. It accomplishes this by organizing these decisions into a hierarchical structure of clusters, criteria, and alternatives. In this approach, the procedural steps are as follows³²:

Criteria identification decision structure: AHP model consists of criteria identification decision structure with goal at first level, clusters at second level and criteria at third level. 

Pairwise Comparisons: A fundamental aspect of AHP is the pairwise comparison process. For each pair of criteria and sub-criteria, individuals or decision-makers are asked to make judgments about their relative importance using saaty’s scale ³². In this study, the pair wise comparison matrix was formed to determine the significance of criteria in relation to the objective. as per Eq. (7). After generating pair wise comparison matrix of clusters and criteria, each element within matrix is normalised by Eq. (8). 

Where, a_ij represent the importance of criterion i over criterion j based on the scale of relative importance.

Computation of priority vector: Each row in the normalized matrix, as per Eq. (9-10), undergoes summation, followed by division by the matrix's size and averaging to derive the weights for each criterion. These resultant weights are designated as priority vectors. 

Calculation of consistency ratio: The consistency ratio (CR) is used to evaluate consistency. To calculate CR, begin by determining the eigenvalue of the matrix (?max) using Eq. (11), then calculate the Consistency Index (CI) with Eq. (12). To assess consistency, knowledge of the 'random index' (RI) is required, and RI values corresponding to each matrix size can be found  as per Satty’s data ³². The consistency ratio is calculated using Eq. (14), and a value below 0.10 indicates a consistent comparison matrix.

Weight Calculation: After collecting the pairwise comparison judgments and passing the consistency check, the calculated weights for each cluster and criterion are employed to compute scores or rankings for each alternative concerning each criterion.

PROMETHEE II decision model 

This decision model offers a structured framework for decision analysis, facilitating informed choices in scenarios with multiple criteria and diverse decision-maker preferences. It is an extension of the original PROMETHEE method, with PROMETHEE-II being preferred for its ability to provide a complete ranking of alternatives.

The processing steps for PROMETHEE II techniques are follows ^33,34:

Prepare decision matrix according Eq. (15).

Where,

A₁, A ₂……Am= numbers of alternatives

X₁₁, X₁₂…..X₁n=values of n number of criteria for alternatives.

The evaluation criteria data are categorized into beneficial and non-beneficial criteria to prioritize the alternatives

Create normalize evaluation matrix according the nature of the selected criteria using Eq. (16) for non-beneficial and Eq. (17) for beneficial criteria. 

The evaluative differences (deviations) are calculated using Eq. (18) by assessing the variances of the i^th alternative compared to the other alternatives.

Evaluating the preference function (Pj) involves using the conditions outlined in Eq. (19-20). If the difference (D) between Ma and Mb is less than or equal to zero, the preference function value is set to zero. Conversely, when D (Ma - Mb) is greater than zero, the preference function value is determined using the differences (R (ij) a ? R (ij) b).  

Calculating the aggregate preference function, ? (Ma, Mb), as per Eq. (21). For this calculation, both direct rating-derived criteria weights and AHP-derived criteria weights were taken into account. This inclusion of direct rating-derived criteria weights is a part of the sensitivity analysis for the proposed hybrid model. This process entails comparing the outcomes of the hybrid AHP-PROMETHEE-II approach with those of the PROMETHEE II decision model, providing insights into the influence of criteria weighting methods on ranking preferences. 

Calculate the surplus and deficit outranking flow of the alternatives using Eq. (22-23) for both PROMETHEE II and AHP-PROMETHEE-II models. The net outranking flow for the alternatives is computed solely for PROMETHEE II and AHP-PROMETHEE II models, enabling the comprehensive ranking of the alternatives. Additionally, calculate the net flow for all options as per Eq. (24). Subsequently, rank the alternatives in descending order based on the net flow (?), thus identifying the best ethanol fuel induction technique for each load range in this study.

To ensure the reliability of comparisons among clusters and criteria in the proposed AHP-PROMETHEE II hybrid model, a consistency ratio (CR) was calculated, demonstrating that the pairwise comparison matrices generated for these clusters and criteria exhibit a high level of consistency by falling below the established threshold of 0.10. This emphasizes the resilience of the model's comparisons. To assess disparity in ranking between two decision models, namely a single PROMETHEE II model and the hybrid AHP-PROMETHEE II model, Spearman's rank coefficient was computed. Resulting value, falling within -1 to +1, indicates a substantial degree of association between rankings generated by these models. Additionally, the results obtained from the proposed AHP-PROMETHEE II model were cross-validated by comparing them with experimental findings reported in the existing literature.

Results and discussion

The research aims to optimize ethanol fuel induction techniques for CI engines across varying engine load ranges, utilizing a combination of diesel-ethanol (DE10) and diesel-ethanol-nanoparticle (DE10_NP) blends, through the hybrid AHP-PROMETHEE II approach. This approach evaluating a total of 18 alternatives, including various combinations of fuel blends (DE10, DE10_NP), ethanol induction methods (blending, fumigation, and combined blending-fumigation), and engine load ranges categorized as low (0 to 20%), medium (40 to 60%), and high (80 to 110%). In this evaluation process, a comprehensive set of criteria-clusters that includes performance, combustion characteristics, emissions, and after-treatment economy were considered. Within each of these cluster, specific sub-criteria were identified, and their corresponding values are derived from the test results. In the following sections, computation and results of our AHP analysis and the application of PROMETHEE II for prioritizing the alternatives were presented.

Computation of AHP model and results

In this section, analytic hierarchy process –AHP is employed to establish criteria weightage essential for achieving goal of selection of best ethanol induction techniques at each load range. Fig. (2) illustrates the decision structuring process, organizing the criteria into four clusters with a total of eighteen criteria for our AHP model.

Figure 2: Criteria identification decision structuring for AHP model   Click here to view Figure

In accordance with Saaty's scale ³², 16 pair-wise comparison of clusters and 86 pair wise comparisons of criteria was formed to establish the significance of clusters and criteria concerning goal. Once generated, these matrices were used to derive normalized matrices for both clusters (Table 6) and criteria (Table 7). Subsequently, computed priority vectors, which provide importance weights for each cluster and criterion. Thus, AHP derived cluster and criteria weight matrix was obtained, as shown in Table 8. For each cluster, a consistency ratio (CR) of 0.05 has been determined, which falls below the established threshold of 0.10. This signifies that the pairwise comparison matrices generated for these clusters exhibit a high level of consistency. Consequently, the calculated weights from the AHP model could be utilised to compute scores and rankings for each alternative concerning each criterion.

Table 6. Normalized clusters matrix 

Table 7: Normalized matrix of criteria.

Table 8: AHP derived cluster and criteria weightage

Computation of PROMETHEE- II and results 

In this section, the PROMETHEE II is used to compute the net outranking flows and to create a unified decision model for ranking alternatives. The data values for the evaluation criteria of ethanol fuel induction techniques were obtained at various engine load ranges have been categorized into beneficial and non-beneficial criteria, as illustrated in Table 9.

Table 9: Beneficial and non beneficial criteria

Further these data values given were normalized as per the nature of the criteria using Eq. (16-17). Subsequently, preference function was derived using the conditions outlined in Eq. (19-20). The values of the preference function matrix were then multiplied with both the direct rating-derived weights (Table 10) and AHP-derived weights (Table 8) to calculate an aggregated preference function matrix. The calculated aggregated preference function matrices for both decision models, PROMETHEE II and the hybrid AHP-PROMETHEE II, are used to calculate the surplus, deficit, and net flows for all alternates in both the PROMETHEE II and AHP-PROMETHEE II models were calculated using Eqs. (22-24). the resulting outranking flows are displayed in Table 11. To determine the ranking of alternatives in this study, the decreasing values of the net outranking flow (?) was considered. This ranking allows to identify the best ethanol fuel induction technique for each load range.

Table 10: Direct rating derived cluster and criteria weightage

Table 11: Outranking flow for PROMETHEE II and hybrid AHP-PROMETHEE II model.

Fig. 3 represent the rankings obtained for all 18 alternatives using the hybrid AHP-PROMETHEE II model with AHP-derived weights and the PROMETHEE II model with direct rating-derived weights.

Figure 3: Priority orders comparison according to PROMETHEE II and hybrid AHP-PROMETHEE II model for all alternatives.   Click here to view Figure

According to the results, the hybrid AHP-PROMETHEE II model ranks the CI engine operated at medium load with ethanol blending, fuelled with (DE10) without the use of nanoparticles as the top choice. Meanwhile, as per the PROMETHEE II model, the engine operated at high load with a combined blending-fumigation approach fuelled with DE10_NP, demonstrates optimal reductions in emissions, performance and combustion enhancements, along with the lowest after-treatment requirements and costs. Additionally, Fig. 4 provides a comprehensive visual representation of the rankings for ethanol fuel induction techniques, categorized separately for low, medium, and high load ranges.

Figure 4: Priority orders comparison according to PROMETHEE II and hybrid AHP-PROMETHEE II model (a) low lad range (b) medium load range and (c) high load range.   Click here to view Figure

Results shows specific recommendations for each load range according to the two models. The results indicate that for low load ranges, the optimum ethanol fuel induction technique is blending fuelled with DE10_NP, as determined by both the hybrid AHP-PROMETHEE II and PROMETHEE II models. In the case of medium load ranges, the choice differs slightly between the models. The PROMETHEE II model favours blending fuelled with DE10_NP with nanoparticles, while the hybrid AHP-PROMETHEE II model recommends blending fuelled with DE10. For high load ranges, both models converge on the same choice, identifying the combined blending-fumigation technique with fuel DE10_NP as the optimum ethanol fuel induction technique. Furthermore, for both models, no deviations in preference orders were observed within the low load range. The largest deviation in preference orders was found in the case of high engine load, while for medium engine load, it was intermediate. 

The results of these models were contrasted with findings from literature involving engine tests incorporating nanoparticles. Findings from these model indicate that optimal results were achieved when nanoparticles were used in conjunction with ethanol blending techniques at low and medium loading condition. However, for high loading condition the most favorable outcomes were obtained with a combination of blending and fumigation. The literature also revealed the same, which indicate that the use of metal-based nano additives in the diesel-ethanol blending emerges as the preferred technique for reducing HC, CO, and NOx emissions. Additionally, this approach showed improvements in engine efficiency, reduced fuel consumption, and enhanced combustion characteristics. Specifically, improvements were observed in terms of heat release rate, delay period, and combustion duration across all engine load conditions ^4,18–24. Furthermore, it was noted in the literature that nanoparticles exhibit catalytic activity in oxidation reactions. When combined with the presence of ethanol in diesel fuel, this catalytic action led to a reduction in particulate matter (PM) and soot emissions. The utilization of metal-based nanoparticles was found to enhance soot oxidation, subsequently decreasing the need for diesel particulate filter (DPF) after-treatment and regeneration processes ^25,26. Additionally, the combined blending-fumigation technique involving ethanol exhibited superior enhancements in engine performance and combustion parameters. Simultaneously, it resulted in a more substantial reduction in emissions when compared to the use of a simple diesel-ethanol blending approach, particularly at high engine loads ¹¹. The results obtained using the proposed AHP-PROMETHEE II model align closely with the experimental findings reported in the existing literature. This similarity validating the utility and reliability of the proposed model.

Spearman’s rank correlation coefficient (p)

Spearman’s rank coefficient '?' serves as the index for quantifying the non-parametric monotonic relationship between two sets of variables, using Eq. (25). In the context of this study, it is employed to assess the effectiveness of the hybrid AHP-PROMETHEE II model by calculating the deviation in rankings compared to the single PROMETHEE II model.  The value of ‘ should lies between -1 to +1. Table 12 presents a comparison of the rankings generated by PROMETHEE II and the hybrid AHP-PROMETHEE II models, along with their corresponding absolute deviations.

 Where,

d_i = pairwise difference between the ranks given by two optimization method.

n = number of variables in associated with the problem. 

It provides insight into the degree of agreement or disparity between the rankings produced by the two decision models. The results indicate that the coefficient (?) between the two models is 0.59 when considering all alternatives.

Table 12: Comparison of ranks obtained by PROMETHEE II and hybrid AHP-PROMETHEE II and their absolute deviation

Conclusion

In this study, a single-cylinder CI engine underwent testing using different fuel blends, including diesel-ethanol (DE10) and diesel-ethanol-nanoparticle (DE10_NP). Various ethanol fuelling techniques were employed, such as blending, ethanol fumigation, and a combination of blending-fumigation. These tests were conducted over a range of engine loads, including low, medium, and high settings. In total, these experimental studies comprised 18 distinct alternatives, which were evaluated using multi criteria decision making hybrid model to identify optimum ethanol fuel induction techniques and fuel blend at each operating load range. The results obtained in the study were given below.

To identify the best ethanol fuel induction technique and fuel blend for each engine load condition, a hybrid decision model was introduced, combining Analytic Hierarchy Process (AHP) with PROMETHEE II. In this model, 18 different alternatives were categorized into four clusters: combustion, emission, performance, and exhaust after-treatment economy. A total of 18 distinct criteria were taken into account for the evaluation. These data were collected from engine test runs that covered three different loading conditions and incorporated six various fuelling techniques and fuel blend combinations.

In hybrid model, AHP model was utilised to compute weights for all four clusters and 18 individual criteria. To accomplish this, a total of 16 pairwise comparisons were conducted for the clusters, and 86 pairwise comparisons were performed for the criteria to establish criteria weights.

AHP incorporates a consistency check through the consistency ratio (CR). For each cluster, a consistency ratio (CR) of 0.05 has been calculated, which is lower than the predefined threshold of 0.10. As a result, the weights obtained from the AHP model can be confidently used to calculate scores and rankings for each alternative with respect to every criterion.

The net flow values for all alternates in both the PROMETHEE II (based on direct rating derived weights) and AHP-PROMETHEE II models (based on AHP-derived weights) were calculated to determine the ranking of alternatives. 

The hybrid AHP-PROMETHEE II model identifies the CI engine operating at medium load with ethanol blending (DE10) and without nanoparticles as the preferred choice. On the other hand, according to the PROMETHEE II model alone, the engine operating at high load with a combined blending-fumigation approach using DE10_NP fuel, demonstrates optimal reductions in emissions, performance and combustion enhancements, along with the lowest after-treatment requirements and costs.

The results of the two models offer specific recommendations for each load range. At low load ranges, both the hybrid AHP-PROMETHEE II and PROMETHEE II models suggest that the optimal ethanol fuel induction technique is blending with DE10_NP. In the case of medium load ranges, PROMETHEE II model recommends blending with DE10_NP with nanoparticles, whereas the hybrid AHP-PROMETHEE II model recommends blending with DE10. However, for high load ranges, both models reach a consensus, pointing to the combined blending-fumigation technique with DE10_NP as the most suitable ethanol fuel induction technique. 

The results indicate that when considering all available alternatives, the spearman's rank coefficient for both models is 0.59. This suggests a significant degree of correlation between the rankings produced by the two models across different engine load conditions. 

Acknowledgements

The author would like to thank Gujarat Technological University and L. D. College of engineering, Ahmedabad, Gujarat for granting the Ph.D. research work. Research assistance laboratory of Apex Innovations Pvt. Ltd. Kupwad, Sangli, Maharashtra, India is highly appreciated for allowed the experimental work.

Conflict of Interest 

The authors declares no conflict of interest.

Funding Sources

There is no funding or financial support for this research work.

References

1. India Bharat Stage vi Emission Standards: International Council on Clean Transportation.; 2016.
2. Sakthivel P, Subramanian KA, Mathai R. Indian scenario of ethanol fuel and its utilization in automotive transportation sector. Resour Conserv Recycl. 2018;132(September 2017):102-120. doi:10.1016/j.resconrec.2018.01.012
   CrossRef
3. Salvi BL, Subramanian KA, Panwar NL. Alternative fuels for transportation vehicles: A technical review. Renew Sustain Energy Rev. 2013;25:404-419. doi:10.1016/j.rser.2013.04.017
   CrossRef
4. Atelge MR. Investigation of a ternary blend of diesel/ethanol/n-butanol with binary nano additives on combustion and emission: A modeling and optimization approach with artificial neural networks. Fuel Process Technol. 2022;229:107155. doi:10.1016/J.FUPROC.2021.107155
   CrossRef
5. Tutak W, Jamrozik A, Pyrc M, Sobiepa?ski M. A comparative study of co-combustion process of diesel-ethanol and biodiesel-ethanol blends in the direct injection diesel engine. Appl Therm Eng. 2017. doi:10.1016/j.applthermaleng.2017.02.029
   CrossRef
6. Gnanamoorthi V, Devaradjane G. Effect of compression ratio on the performance, combustion and emission of di diesel engine fueled with ethanol e Diesel blend. J Energy Inst. 2015;88(1):19-26. doi:10.1016/j.joei.2014.06.001
   CrossRef
7. Armas O, García-Contreras R, Ramos Á. Pollutant emissions from engine starting with ethanol and butanol diesel blends. Fuel Process Technol. 2012;100:63-73. doi:10.1016/j.fuproc.2012.03.003
   CrossRef
8. Britto RF, Martins CA. Experimental analysis of a diesel engine operating in Diesel-Ethanol Dual-Fuel mode. Fuel. 2014;134(May):140-150. doi:10.1016/j.fuel.2014.05.010
   CrossRef
9. Murcak A, Ha?imo?lu C, Çevik I, Kahraman H. Effect of injection timing to performance of a diesel engine fuelled with different diesel-ethanol mixtures. Fuel. 2015;153:569-577. doi:10.1016/j.fuel.2015.03.028
   CrossRef
10. Sarjovaara T, Alantie J, Larmi M. Ethanol dual-fuel combustion concept on heavy duty engine. Energy. 2013. doi:10.1016/j.energy.2013.10.053
    CrossRef
11. Ghadikolaei MA, Cheung CS, Yung KF. Study of combustion, performance and emissions of a diesel engine fueled with ternary fuel in blended and fumigation modes. Fuel. 2019;235(June 2018):288-300. doi:10.1016/j.fuel.2018.07.089
    CrossRef
12.  Jamrozik A, Tutak W, Pyrc M, Gruca M, Ko?iško M. Study on co-combustion of diesel fuel with oxygenated alcohols in a compression ignition dual-fuel engine. Fuel. 2018;221(September 2017):329-345. doi:10.1016/j.fuel.2018.02.098
    CrossRef
13.  Pedrozo VB, May I, Guan W, Zhao H. High efficiency ethanol-diesel dual-fuel combustion: A comparison against conventional diesel combustion from low to full engine load. Fuel. 2018;230(May):440-451. doi:10.1016/j.fuel.2018.05.034
    CrossRef
14. Lee J, Lee S, Lee S. Experimental investigation on the performance and emissions characteristics of ethanol/diesel dual-fuel combustion. Fuel. 2018;220(December 2017):72-79. doi:10.1016/j.fuel.2018.02.002
    CrossRef
15.  Jamuwa DK, Sharma D, Soni SL. Experimental investigation of performance, exhaust emission and combustion parameters of stationary compression ignition engine using ethanol fumigation in dual fuel mode. Energy Convers Manag. 2016;115:221-231. doi:10.1016/j.enconman.2016.02.055
    CrossRef
16. Pedrozo VB, May I, Dalla Nora M, Cairns A, Zhao H. Experimental analysis of ethanol dual-fuel combustion in a heavy-duty diesel engine: An optimisation at low load. Appl Energy. 2016;165:166-182. doi:10.1016/j.apenergy.2015.12.052
    CrossRef
17. Padala S, Woo C, Kook S, Hawkes ER. Ethanol utilisation in a diesel engine using dual-fuelling technology. Fuel. 2013. doi:10.1016/j.fuel.2013.03.049
    CrossRef
18. Wu Q, Xie X, Wang Y, Roskilly T. E ff ect of carbon coated aluminum nanoparticles as additive to biodiesel- diesel blends on performance and emission characteristics of diesel engine ?. Appl Energy. 2018;221(March):597-604. doi:10.1016/j.apenergy.2018.03.157
    CrossRef
19. Nour M, EL-Seesy AI, Abdel-Rahman AK, Bady M. Influence of adding aluminum oxide nanoparticles to diesterol blends on the combustion and exhaust emission characteristics of a diesel engine. Exp Therm Fluid Sci. 2018;98:634-644. doi:10.1016/j.expthermflusci.2018.07.009
    CrossRef
20. Prabakaran B, Udhoji A. Experimental investigation into effects of addition of zinc oxide on performance , combustion and emission characteristics of diesel-biodiesel-ethanol blends in CI engine. Alexandria Eng J. 2016. doi:10.1016/j.aej.2016.08.022
    CrossRef
21. Venu H, Raju VD, Lingesan S, Elahi M Soudagar M. Influence of Al2O3nano Additives in Ternary Fuel (Diesel-Biodiesel-Ethanol) Blends Operated in a Single Cylinder Diesel Engine: Performance, Combustion and Emission Characteristics. Vol 215. Elsevier Ltd; 2021. doi:10.1016/j.energy.2020.119091
    CrossRef
22. Atelge MR, Arslan E, Kahraman N, Ünalan S. Evaluation of hybrid nanoparticles to oxygenated fuel with ethanol and n- butanol on combustion behavior. Fuel. 2023;344. doi:10.1016/j.fuel.2023.128048
    CrossRef
23. Mohapatra T, Mishra SS, Sahoo SS, Albani A, Awad MM. Performance, emissions, and economic evaluation of a VCR CI engine using a bio-ethanol and diesel fuel combination with Al2O3 nanoparticles. Case Stud Therm Eng. 2023;49(July):103375. doi:10.1016/j.csite.2023.103375
    CrossRef
24. Vijayan S, Sathyamurthy R, Mokheimer EMA, Kumar RS. Performance enhancement and emission reduction of CRDI diesel engine fueled using Manilkara Zapota biodiesel blend with TiO2 nanoadditive. Fuel Process Technol. 2023;248:107842. doi:10.1016/J.FUPROC.2023.107842
    CrossRef
25. Mohankumar S, Senthilkumar P. Particulate matter formation and its control methodologies for diesel engine: A comprehensive review. Renew Sustain Energy Rev. 2017;80:1227-1238. doi:10.1016/J.RSER.2017.05.133
    CrossRef
26. Dhahad HA, Hasan AM, Chaichan MT, Kazem HA. Prognostic of diesel engine emissions and performance based on an intelligent technique for nanoparticle additives. Energy. 2022;238:121855. doi:10.1016/j.energy.2021.121855
    CrossRef
27. Erdo?an S, Balki MK, Ayd?n S, Sayin C. The best fuel selection with hybrid multiple-criteria decision making approaches in a CI engine fueled with their blends and pure biodiesels produced from different sources. Renew Energy. 2019;134:653-668. doi:10.1016/j.renene.2018.11.060
    CrossRef
28. Suryabhanji Bhojane M, Chandra Murmu S, Chattopadhyay H, Dutta A. Application of MCDM technique for selection of fuel in power plant. Mater Today Proc. November 2023. doi:10.1016/J.MATPR.2023.10.159
    CrossRef
29. Sakthivel G, Ilangkumaran M. A hybrid multi-criteria decision making approach of ANP and TOPSIS to evaluate the optimum fuel blend in IC engine. Int J Decis Support Syst. 2015;1(3):268. doi:10.1504/ijdss.2015.070170
    CrossRef
30. Erdo?an S, Ayd?n S, Balki MK, Sayin C. Operational evaluation of thermal barrier coated diesel engine fueled with biodiesel/diesel blend by using MCDM method base on engine performance, emission and combustion characteristics. Renew Energy. 2020;151:698-706. doi:10.1016/j.renene.2019.11.075
    CrossRef
31. Balki MK, Erdo?an S, Ayd?n S, Sayin C. The optimization of engine operating parameters via SWARA and ARAS hybrid method in a small SI engine using alternative fuels. J Clean Prod. 2020;258. doi:10.1016/j.jclepro.2020.120685
    CrossRef
32. Saaty RW. The analytic hierarchy process-what it is and how it is used. Math Model. 1987;9(3-5):161-176. doi:10.1016/0270-0255(87)90473-8
    CrossRef
33. Goswami SS. Outranking Methods: Promethee i and Promethee II. Found Manag. 2020;12(1):93-110. doi:10.2478/fman-2020-0008
    CrossRef
34. Sakthivel G, Ilangkumaran M, Nagarajan G, Raja A, Ragunadhan PM, Prakash J. A hybrid MCDM approach for evaluating an automobile purchase model. Int J Inf Decis Sci. 2013;5(1):50-85. doi:10.1504/IJIDS.2013.052017
    CrossRef