The Impact of Oil Viscosity and Fuel Quality on Internal Combustion Engine Performance and Emissions: An Experimental Approach

Abstract

The automotive industry faces increasing challenges due to fuel scarcity and pollutant emissions, necessitating the implementation of strategies that optimize engine performance while minimizing the environmental impact. This study aimed to analyze the influence of oil viscosity and fuel quality on the engine performance and pollutant emissions in an internal combustion engine. A Response Surface Methodology (RSM)-based experimental design was employed. Three oil viscosity levels (SAE 5W-30, 10W-30, and 20W-50) and three fuel quality levels (87, 92, and 95 octane) were evaluated using a Chevrolet Grand Vitara 2.0L (General Motors, Quito, Ecuador) tested on a dynamometer. The oil grades were selected to represent a practical range of viscosities commonly used in commercial vehicles operating under local conditions. The results indicate that using lower-viscosity oil (SAE 5W-30) increased the engine power by up to 6.25% compared to when using SAE 20W-50. Additionally, using higher-octane fuel led to an average power increase of 1.49%, attributed to improved combustion stability and the ability to operate at a more advanced ignition timing without knocking. The emissions analysis revealed that high-viscosity oil at high RPMs increased CO₂ emissions to 14.4% vol, whereas low-viscosity oil at low RPMs reduced CO₂ emissions to 13.4% vol. Statistical analysis confirmed that the engine speed (RPM) was the most influential factor in emissions (F = 163.11 and p < 0.0001 for CO₂; F = 247.02 and p < 0.0001 for NOx), while fuel quality also played a significant role. These findings suggest that optimizing the oil viscosity and selecting the appropriate fuel can enhance engine efficiency and reduce emissions, thereby contributing to the development of more sustainable automotive technologies. Future research should explore the use of ultra-low-viscosity lubricants (SAE 0W-20) and assess their long-term effects on engine wear.

1. Introduction

The study of internal combustion engine (ICE) behavior remains a key area of research in the development of more efficient and sustainable technologies. Despite the growing adoption of electric and hybrid systems, conventional engines continue to dominate a significant portion of the global vehicle fleet, making it essential to further optimize their performance. In particular, analyzing variables such as the lubricating oil viscosity, fuel quality, and engine operating conditions and applying statistical techniques like the Response Surface Methodology (RSM) are critical to reducing energy consumption and minimizing pollutant emissions without compromising power and torque outputs.

One of the most influential factors in engine performance is fuel quality, particularly the fuel’s octane rating. Rodríguez-Fernández et al. reported that higher-octane fuels enable engine configurations to be optimized, improving thermal efficiency and increasing the power without knocking, as they allow for the use of higher compression ratios [1]. Similarly, Fernanda et al. demonstrated that high-octane fuels reduce specific fuel consumption and enhance engine acceleration [2]. However, several studies have warned that using fuels with an octane number above the engine’s design requirements can have adverse effects. Sayin observed that in engines not calibrated for high-octane fuels, the brake thermal efficiency decreases while fuel consumption increases [3], and Sayın et al., along with Athafah et al., found increased pollutant emissions without significant gains in the engine power [4,5]. Regarding emissions, Al-Farayedhi and Fernanda et al. concluded that higher-octane fuels promote more complete combustion, reducing carbon monoxide (CO) and hydrocarbon (HC) emissions [2,6]. However, Iodice et al. reported that the higher combustion temperatures associated with these fuels can increase nitrogen oxide (NOx) emissions [7]. Moreover, Khalifa et al. warned that excessive octane in engines lacking knock control systems can increase the total hydrocarbon (THC) emissions without offering meaningful power benefits [8].

Lubricating oils—especially low-viscosity formulations—have also proven to be a key factor in enhancing thermal efficiency and reducing fuel consumption in internal combustion engines (ICEs). Zhang et al. reported that using 0W-16 oil can increase mechanical efficiency by 3.3% and improve fuel economy by up to 2.8% compared to conventional 5W-30 oils [9]. However, this advantage comes with increased wear due to asperity contact, particularly in critical components like the compression ring. Hei et al. observed a 25% increase in localized wear under high-load conditions when using low-viscosity oils [10]. To address this issue, the use of functional additives—such as viscosity index improvers, anti-wear agents, and pour point depressants—is essential. Mu et al. demonstrated that these additives can enhance wear resistance by up to 20% [11]. Additionally, Palikhel et al. found that blends of natural and synthetic oils improve engine performance and reduce frictional losses, resulting in 3–5% fuel savings compared to conventional oils [12]. However, Giakoumis noted that high-viscosity oils can increase frictional losses during transient engine operation, reducing the dynamic response and causing power losses of up to 6% in turbocharged diesel engines [13]. Furthermore, oil viscosity degradation over time has been associated with a 15% increase in engine vibration and a reduced engine lifespan, according to Santana [14].

The engine speed (RPM) is another crucial factor that affects combustion stability, fuel consumption, and thermal efficiency. Ayele et al. found that in DM-TJI systems, smaller orifices improve efficiency at lower speeds (1500–2500 RPM), while larger orifices perform better at higher speeds (3000–4000 RPM) [15]. Song et al. demonstrated that optimizing the speed profile in spark ignition engines can achieve up to 40.5% net indicated efficiency in almost knock-limited conditions [16]. Olaiya et al. reported that in a two-stroke diesel engine, 2000 RPM offers the best balance between power and fuel economy [17]. Similarly, Shinde et al. found that gasoline engines with digital ignition systems reached a maximum brake thermal efficiency of 32% at 6000 RPM [18]. In hydrogen-fueled turbocharged engines, Wang et al. observed thermal efficiencies of above 38% at specific optimized speeds [19]. However, Ghaderi Masouleh et al. showed that increasing the engine speed from 1500 to 3500 RPM results in up to a 30% rise in cycle-to-cycle variations, compromising combustion stability [20]. Additionally, Gao et al. reported that an increased engine speed raises pumping losses and in-cylinder pressure dynamics, negatively affecting the overall intake efficiency [21].

The Response Surface Methodology (RSM) is a robust set of statistical and mathematical techniques used to model and analyze systems in which multiple variables simultaneously influence one or more responses of interest. By enabling the development of empirical models, the RSM facilitates the identification of significant factors, the quantification of their interaction effects, and the optimization of system performance. In the context of engine research, the RSM has proven to be a practical and efficient approach for evaluating complex interactions between different variables. Ijaz Malik et al. reported a 7.2% improvement in the brake power and a 9.5% reduction in CO emissions in SI engines fueled with methanol–gasoline blends under RSM-optimized conditions [22]. Safieddin Ardebili et al. used the RSM in the optimization of HCCI engines powered by fusel oil and diethyl ether blends, achieving a 5.2% increase in thermal efficiency [23]. Similarly, Katekaew et al. optimized a single-cylinder diesel engine running on biodiesel from waste cooking oil, reporting a 12% reduction in HC emissions and a 6.5% increase in the brake thermal efficiency [24]. Liang et al. and Kumar et al. also found that the RSM-enabled optimization of the injection pressure and fuel composition led to over a 10% improvement in performance with significant emissions reductions in diesel engines [25,26].

To further contextualize the current study and highlight the methodological distinctions in the existing body of research,

Table 1. Overview of selected studies on engine performance and emissions.

Author(s)	Engine Type	Variables Studied	Methodology
Zhang et al. [9]	Gasoline engine	Oil viscosity (0W-16 vs. 5W-30)	Experimental evaluation of fuel economy and wear
Fernanda et al. [2]	Spark ignition engine	Octane rating effect	Performance tests and measurement of fuel consumption
Sayın et al. [4]	Gasoline engine (non-optimized)	High-octane fuel in standard engine	Comparative fuel analysis
Iodice et al. [7]	Spark ignition engine	Octane rating and NOx formation	Combustion analysis
Giakoumis [13]	Turbocharged diesel engine	High-viscosity oil and transient operation	Simulation and empirical data
Khalifa et al. [8]	Engine without knock control	Octane rating and THC emissions	Experimental tests

summarizes selected studies that have examined the effects of oil viscosity, fuel quality, and engine operation parameters on performance and emissions in internal combustion engines. These studies varied in terms of the engine type, experimental conditions, and methodological focus, ranging from simulation-based analyses to empirical testing under controlled or idealized environments.

The recent literature highlights a lack of comprehensive studies that concurrently evaluate the combined influence of the oil viscosity, fuel octane rating, and engine speed on critical performance indicators such as the power, torque, and exhaust emissions. While previous research has explored these variables independently, particularly their individual effects on engine performance and environmental impacts, there remains a notable absence of experimental approaches that integrate all three under realistic vehicle operating conditions using commercially available fuels and lubricants. Most prior investigations have either focused on isolated parameters or employed controlled laboratory simulations that do not fully reflect on-road dynamics.

This study addressed that gap by applying the Response Surface Methodology (RSM) to systematically investigate the interaction between the lubricant viscosity, fuel quality, and rotational speed in a Chevrolet Grand Vitara 2008, tested on a chassis dynamometer. Using locally sourced fuels and lubricants standard in the Ecuadorian market added further relevance to the analysis, offering practical insight into how these variables jointly affect combustion efficiency and emission profiles in conventional internal combustion engines. This approach also supports decision-making in regional contexts where fuel standards and lubricant options vary significantly.

The remainder of this paper is organized as follows: Section 2 details the experimental materials and procedures; Section 3 presents the development and results of the statistical model; Section 4 discusses the findings in light of the existing literature; and Section 5 outlines the main conclusions along with suggested directions for future research.

2. Materials and Methods

2.1. Experimental Setup Description

This research’s data collection was structured using a protocol aligned with the Response Surface Methodology (RSM). This framework systematically examined how the engine oil viscosity, fuel quality (represented by the octane rating), and engine speed (RPM) influenced the engine’s performance and pollutant emission behavior.

The experiments were conducted using a Chevrolet Grand Vitara featuring a 2.0 L SOHC, four-stroke, four-cylinder spark ignition (SI) gasoline engine. Detailed engine specifications are listed in

Table 2. Main characteristics of the test engine.

Technical Specifications
Engine	2.0 L SOHC
Valves	16
Number of cylinders	4
Power (CV @ rpm)	140 @ 5600
Torque (Nm @ rpm)	183 @ 4000
Fuel supply	Multi-point fuel injection (MPFI)
Compression ratio	10.5:1
Final ratio	3.944
Gross vehicle weight	1608 Kg

. The vehicle was tested on an MAHA LPS 3000 chassis dynamometer (MAHA Maschinenbau Haldenwang GmbH & Co. KG, Haldenwang, Germany), which incorporated eddy current braking technology. This equipment is capable of measuring the engine traction and power. It can apply variable loads at rotational speeds ranging from 0 to 10,000 rpm, simulating speeds of up to 260 km/h, with a constant tractive force capacity of up to 6 kN, as illustrated in Figure 1.

The exhaust gases were measured using a Brain Bee gas analyzer (Mahle, Stuttgart, Germany), which featured a high-accuracy detection chamber. This equipment provides consistent and reliable emission data, ensuring reproducibility across trials. The technical characteristics of the analyzer are detailed in

Table 3. Technical specifications of the gas analyzer.

Measuring Fields	Range	Unit	Resolution
CO	0–9.99	% vol	0.01
CO2	0–19.9	% vol	0.1
HC hexane	0–9999	ppm vol	1
O2	0–25	% vol	0.01
NOx	0–5000	ppm vol	1
Revolution inductance/capacitance	300–9990	rpm	10
Oil temperature	20–150	°C	1

.

The lubricants employed in this study, SAE 5W-30, 10W-30, and 20W-50, represented a practical spectrum of viscosity grades commonly used in light-duty vehicles within Ecuador. Their selection was guided by the local market availability and intended to simulate real-world conditions by comparing low-, mid-, and high-viscosity oils. The input variable, the oil viscosity, was defined according to the kinematic viscosity measured at 100 °C (in mm²/s). All three lubricants were primarily formulated using synthetic polyalphaolefin (PAO) bases, comprising 70–90% of the total composition. Synthetic esters were blended in 10–20% proportions to enhance lubrication characteristics. Each formulation included functional additive packages, comprising 10–20% of the oil volume. These packages contained detergents (1–3%), dispersants (3–5%), antioxidants (1–2%), corrosion inhibitors (0.5–1%), viscosity modifiers (2–5%), friction-reducing agents (0.1–1%), anti-wear components (1–2%), and foam suppressants (<0.1%).

Table 4. Technical specifications of the oils.

SAE Grade	5W30	10W30	20W50
Specific gravity @ 15 °C	0.861	0.866	0.881
Density, g/mL @ 15 °C	0.859	0.864	0.878
Color, ASTM D1500	3.0	3.0	3.0
Flash point (COC), °C (°F)	216 (421)	229 (444)	230 (446)
Pour point, °C (°F)	−39 (−38)	−39 (−38)	−30 (−22)
Kinematic viscosity, mm2/s @ 40 °C	66.2	65.7	176
Kinematic viscosity, mm2/s @ 100 °C	9.66	14.08	18.5
Viscosity index	158	148	128
Cold cranking viscosity, cP @ (°C)	6150 (−30)	4550 (−25)	7200 (−15)
High-temp/high-shear viscosity, cP @ 150 °C	3.1	3.0	4.9

provides detailed specifications of the lubricants tested.

The variable fuel quality was established based on the octane ratings of Ecuador’s commercially available fuels. Two primary gasoline types dominate the market: Ecopaís gasoline, with an octane number of 87, and Super gasoline, rated at 92 octane.

Ecopaís gasoline is a biofuel comprising 5% bioethanol and 95% base gasoline. Its production requires 62% high-octane naphtha (NAO), 33% low-octane naphtha (NBO), and 5% bioethanol. Due to the high-octane rating of bioethanol, the production of NAO is reduced by 14%, thereby decreasing the dependency on imported derivatives and generating a positive economic impact on the national economy. On the other hand, Super gasoline is used in high-compression-ratio engines. Its formulation includes a mixture of hydrocarbons, mainly isoparaffins, and aromatics, which enhance the fuel’s resistance to high pressures and temperatures, preventing molecular degradation.

According to the NTE INEN 935 standard [27], fuels used in internal combustion engines must comply with specific requirements.

Table 5. Gasoline requirements in Ecuador.

Requirements	Ecopaís Gasoline	Super Gasoline
Research octane number (RON)	87	92
Vapor pressure [kPa]	60	60
Distillation residue [%]	2 max	2 max
Gum content [mg/100 mL]	3 max	4 mx
Sulfur content [%]	0.065 max	0.065 max
Aromatics content [%]	30 max	30 max
Benzene content [%]	1 max	2 mx
Olefin content [%]	18 max	25 max
Oxygen content [%	2.7 max	2.7 max
Pb content [mg/L]	Not detectable	Not detectable
Manganese content [mg/L]	Not detectable	Not detectable
Iron content [mg/L]	Not detectable	Not detectable

presents a comparison of the specifications of Ecopaís and Super gasoline.

To define the levels of this variable in the study, an octane booster was used, an additive that rapidly and cost-effectively enhances a fuel’s chemical properties. This approach enabled the evaluation of the effects of higher-octane fuel without requiring complex reformulations. Based on this method, the following fuel quality levels were established:

• Lower level: Ecopaís gasoline (−1).
• Medium level: Super gasoline (0).
• Upper level: Super gasoline with an octane booster additive (1).

2.2. Response Surface Methodology

The Response Surface Methodology (RSM) is a statistical technique widely used for modeling and analyzing processes in which several independent variables influence a dependent variable or response. In this work, where the number of variables $k=3$, including the oil viscosity, fuel quality, and engine speed, the method was used to identify and quantify their effect on emissions and engine performance. Although the proper functional form relating these inputs to the output is unknown, it can be approximated using a low-degree polynomial equation, as shown in Equation (1) [28]:

$$y=f\prime \left(x\right)\beta +\epsilon$$

(1)

where ${x=(x}_1,x_2, \dots , x_k)\prime$ denotes the vector of independent variables, and $f\left(x\right)$ represents a vector function including linear and higher-order terms and interactions among the variables. The coefficient vector $\beta$ contains the unknown parameters to be estimated, and $\epsilon$ is a random error assumed to have a mean of zero. When Equation (1) adequately represents the process, $f\prime \left(x\right)\beta$ estimates the expected response $\mu \left(x\right)$. In practice, the RSM commonly employs two polynomial models derived from Equation (1). The first-order model, Equation (2), assumes a linear relationship between the response and each input [29]:

$$y={\beta }_0+\sum _{i=1}^k{\beta }_i{X}_i+\epsilon$$

(2)

where $X_i$ represents the $i$th factor and ${\beta }_i$ its corresponding coefficient. This model does not account for curvature or interaction effects. To capture nonlinearities and interactions, a second-order polynomial model is used, as expressed in Equation (3) [28]:

$$y={\beta }_0+\sum _{i=1}^k{\beta }_i{x}_i+\sum _{i=1}^k \sum _{j\ge 1}^k{\beta }_{ij}{x}_i{x}_j+\sum _{j=1}^k{\beta }_{ii}{x_i}^2+\epsilon$$

(3)

where the coefficient ${\beta }_{ij}$ represents the interaction effects between input variables, and ${\beta }_{ii}$ accounts for quadratic curvature. This model enables a more comprehensive analysis of how the variables jointly influence the output.

For this study, a Box–Behnken design was employed, which is known for providing efficient and precise estimates of the variables’ effects on the response. This design is particularly suitable for modeling quadratic relationships, requiring fewer runs than full factorial designs. In this case, three factors at three levels were evaluated, resulting in 15 unique experimental combinations with 3 replicates per condition, totaling 45 trials.

Table 6. Input variables and levels for the experiment.

Factor	Unit	Lower Level	Middle Level	Upper Level
Viscosity	mm2/s	9.66	14.8	18.5
Fuel quality	Research octane number (RON)	−1	0	1
Engine speed	rpm	750	1500	2500

and

Table 7. Table of the design used to achieve the experimental results.

Std Order	Run Order	Pt Type	Blocks	Viscosity (mm2/s)	rpm	Fuel (RON)	CO2 (% vol)	HC (ppm)	NOx (ppm)
30	1	0	2	14.8	1500	0	13.7	23	1.09
18	2	2	2	9.66	2500	0	14.4	1	1.023
16	3	2	2	9.66	750	0	13.5	1	1.09
26	4	2	2	14.8	750	1	12.9	30	1.127
23	5	2	2	18.5	1500	1	14.2	9	1.054
21	6	2	2	18.5	1500	−1	14.2	11	1.057
29	7	0	2	14.8	1500	0	14.2	33	1.07
27	8	2	2	14.8	2500	1	14	22	1.032
22	9	2	2	9.66	1500	1	14.2	8	1.05
25	10	2	2	14.8	2500	−1	13.9	10	1.05
24	11	2	2	14.8	750	−1	13.6	12	1.09
20	12	2	2	9.66	1500	−1	14.1	26	1.042
19	13	2	2	18.5	2500	0	14.5	1	1.024
28	14	0	2	14.8	1500	0	13.9	32	1.08
17	15	2	2	18.5	750	0	13.5	1	1.102
12	16	2	1	14.8	2500	1	14.3	23	1.019
15	17	0	1	14.8	1500	0	13.9	23	1.076
8	18	2	1	18.5	1500	1	14.4	9	1.051
13	19	0	1	14.8	1500	0	13.9	23	1.079
14	20	0	1	14.8	1500	0	14.2	21	1.05
10	21	2	1	14.8	2500	−1	13.9	10	1.07
11	22	2	1	14.8	750	1	13.1	31	1.113
4	23	2	1	18.5	2500	0	14.5	2	1.025
3	24	2	1	9.66	2500	0	14.4	1	1.024
9	25	2	1	14.8	750	−1	13	13	1.1
6	26	2	1	18.5	1500	−1	14.2	10	1.044
2	27	2	1	18.5	750	0	13.6	2	1.088
7	28	2	1	9.66	1500	1	14.3	8	1.046
5	29	2	1	9.66	1500	−1	14.1	25	1.04
1	30	2	1	9.66	750	0	13.5	1	1.09
41	31	2	3	14.8	750	1	13.3	31	1.104
35	32	2	3	9.66	1500	−1	14.2	26	1.038
32	33	2	3	18.5	750	0	13.6	1	1.09
37	34	2	3	9.66	1500	1	14.3	7	1.05
39	35	2	3	14.8	750	−1	13.5	13	1.08
36	36	2	3	18.5	1500	−1	14.2	11	1.045
45	37	0	3	14.8	1500	0	14.2	22	1.068
33	38	2	3	9.66	2500	0	14.4	1	1.023
38	39	2	3	18.5	1500	1	14.3	8	1.063
42	40	2	3	14.8	2500	1	14.3	22	1.023
43	41	0	3	14.8	1500	0	14.2	32	1.05
44	42	0	3	14.8	1500	0	14	33	1.074
34	43	2	3	18.5	2500	0	14.3	2	1.026
40	44	2	3	14.8	2500	−1	13.5	11	1.08
31	45	2	3	9.66	750	0	13.6	2	1.085

present, respectively, the levels of the input variables and the complete experimental design.

3. Results

3.1. Model for CO₂

The construction of the mathematical model for estimating CO₂ emissions began with identifying critical parameters that defined a robust baseline.

Table 8. ANOVA results for CO₂.

Source	Sum of Squares	Df	Mean Square	F-Value	p-Value
Model	6.72	9	0.7465	31.40	<0.0001	Significant
A—Viscosity	0.0099	1	0.0099	0.4183	0.5220
B—RPM	3.88	1	3.88	163.11	<0.0001
C—Fuel	0.0938	1	0.0938	3.94	0.0549
AB	0.0003	1	0.0003	0.0141	0.9062
AC	0.0016	1	0.0016	0.0681	0.7956
BC	0.3594	1	0.3594	15.11	0.0004
A2	0.9348	1	0.9348	39.31	<0.0001
B2	1.63	1	1.63	68.59	<0.0001
C2	0.0820	1	0.0820	3.45	0.0717
Residual	0.8322	35	0.0238
Lack of fit	0.0300	3	0.0100	0.3992	0.7545	Not significant
Pure error	0.8022	32	0.0251
Cor total	7.55	44

summarizes the results of the ANOVA and model fit indicators, offering a clear assessment of the model’s reliability and the significance of the selected factors.

According to the ANOVA results for the quadratic model, the analysis confirmed that the model was statistically significant (F = 31.40, p < 0.0001), indicating a strong influence of the independent variables on the CO₂ outcome. Among them, the engine speed (factor B) was the dominant variable (F = 163.11, p < 0.0001), highlighting its primary contribution to the response. The interaction between the fuel type and engine speed (BC) was also statistically relevant (F = 15.11, p = 0.0004), suggesting a synergistic effect. Additionally, the quadratic components of the viscosity (A²) and RPM (B²) demonstrated significant nonlinear influence, with F-values of 39.31 and 68.59, respectively.

On the other hand, the linear effect of the viscosity (A) and its interaction with the fuel (AC) did not reach significance (p > 0.52 and p > 0.79). While the quadratic term for the fuel (C²) showed a near-significant trend (p = 0.0717), it remained above the conventional threshold. The coefficient of determination (R²) was calculated as 0.8898 (refer to

Table 9. Coefficient of determination for CO₂.

Coefficient of Determination	Value
R2	0.8898
Adjusted R2	0.8614
Predicted R2	0.8236
Adeq Precision	18.0130

), indicating that the model captured approximately 88.98% of the variance. The adjusted R² value of 0.8614 reflects a slight correction when accounting for the model complexity, and the predicted R² of 0.8236 shows consistency between the model prediction and experimental data, affirming the model’s suitability for generalization.

Although the oil viscosity is generally recognized as a key contributor to hydrodynamic friction and therefore to engine efficiency and CO₂ emissions, in this study it did not show a statistically significant influence in the CO₂ model. This result may be attributed to the relatively narrow range of viscosity values examined (5W30 to 20W50), which, under the dynamometer tests’ specific operating conditions and temperature range, may not have been sufficient to produce a distinguishable effect on CO₂ emissions. Furthermore, the dominant effect of the engine speed on combustion dynamics could have overshadowed the influence of the viscosity within this experimental window. Nevertheless, this does not invalidate the importance of the oil viscosity in engine performance, as supported by its significant effect observed in the models for other emissions and performance variables discussed later in this study.

The lack of fit was found to be statistically non-significant, with an F-value of 0.40 and a p-value of 0.7545. This suggests that the model sufficiently explained the variability in the data, eliminating the need for a more complex formulation. The Adequate Precision index was 18.013, indicating a favorable signal-to-noise ratio. Since values above 4 are typically recommended to ensure adequate discrimination between model predictions, this result confirms the model’s robustness and suitability for navigating the response surface.

The final regression model estimating CO₂ emissions using actual factor values is presented in Equation (4). This formulation allows for the prediction of the CO₂ output for given values of the viscosity, RPM, and fuel type, provided these inputs are in their natural units. It is important to note, however, that the scaling of each variable influences the coefficients in this equation, and the constant term does not correspond to the center point of the experimental design. Therefore, the model should not be interpreted as determining the relative effects of individual variables.

CO₂ [% vol] = 14.92 − 0.43 × Viscosity + 0.0022 × RPM − 0.22 × Fuel − 1.35 × 10⁻⁶ × Viscosity × RPM − 0.0026 × Viscosity × Fuel + 0.0002 × RPM × Fuel + 0.015 × Viscosity² − 5.14 × 10⁻⁷ × RPM² − 0.086 × Fuel²

(4)

Additionally, the contour plot (see Figure 2a) illustrates how the oil viscosity (ranging from 9.66 to 18.5 mm²/s) and engine speed (750 to 2500 rpm) jointly influenced CO₂ emission levels. The highest CO₂ concentrations (~14.4% vol) were observed under conditions of high viscosity (18.5 mm²/s) and an elevated engine speed (2500 rpm). This trend suggests that increased viscosity induces greater internal friction, raising the load on the engine and thereby increasing fuel consumption, leading to higher CO₂ emissions. Additionally, incomplete combustion events may be reduced at a higher rpm, enhancing CO₂ production due to the more complete oxidation of hydrocarbons.

The response surface plot (Figure 2b) reinforces these observations, showing that CO₂ emissions peaked at intermediate-to-high viscosity levels and an increased engine speed. At low viscosities (~9.66 mm²/s) and a low rpm (~750), CO₂ emissions were reduced (~13.4% vol). However, as both the viscosity and RPM increased, CO₂ emissions tended to rise, with the maximum values approaching ~14.2% vol. Although the difference in emissions between the lowest (9.66 mm²/s) and highest (18.5 mm²/s) viscosities was modest, this trend still reflects the contribution of increased hydrodynamic resistance and combustion loads to fuel consumption. The results underline that even small variations in the lubricant viscosity can influence CO₂ formation, especially when combined with high engine speeds.

The analysis of the interaction between the oil viscosity and fuel type (octane rating) shown in Figure 3 revealed that, like the viscosity and rpm interaction, a higher oil viscosity tended to increase CO₂ concentrations, which reached up to 14.3% vol when combined with higher-octane fuels. However, unlike the influence of the rpm, where an increased engine speed significantly elevated emissions, the impact of the fuel type was more moderate, with a minor variation in the CO₂ concentration observed when using lower-viscosity oils and lower-octane fuels, reducing emissions to around 14.0% vol. This suggests that while both factors are essential, the engine speed exerts a more pronounced effect on emissions than the fuel type.

Finally, the interaction between the fuel type (octane rating) and engine speed (rpm) shown in Figure 4 indicates that, like in the viscosity analyses, an increase in the rpm significantly raised CO₂ concentrations, which reached up to 14.2% vol at 2500 rpm with higher-octane fuels. However, compared to the interaction between the viscosity and rpm, the influence of the fuel type was more moderate, exerting a lesser impact on CO₂ concentrations when lower-octane fuels were used, which reduced emissions to around 13.2% vol at a lower rpm. This observation reinforces that the engine speed is a more decisive factor affecting CO₂ emissions than the fuel type, although both aspects remain critical for optimizing emissions.

In conclusion, the experimental findings confirm that while the engine speed is the main driver of CO₂ emissions, the fuel type and lubricant viscosity also contribute to combustion efficiency. Although the overall effect of the viscosity was not statistically significant in the linear term, its influence was evident in the nonlinear behavior of the system. This suggests that optimal combinations of operating parameters can still enhance environmental performance without compromising the engine output. These insights, derived from controlled dynamometer tests using commercially available fuels and oils, offer practical relevance for regions with similar energy and regulatory contexts.

3.2. Model for HCs

The HC emissions model, derived through quadratic regression and evaluated using an ANOVA (

Table 10. ANOVA results for HCs.

Source	Sum of Squares	Df	Mean Square	F-Value	p-Value
Model	4175.25	9	463.92	13.37	<0.0001
A—Viscosity	64.85	1	64.85	1.87	0.1803
B—RPM	43.19	1	43.19	1.24	0.2722
C—Fuel	14.97	1	14.97	0.4315	0.5156
AB	0.5239	1	0.5239	0.0151	0.9029
AC	299.09	1	299.09	8.62	0.0058
BC	5.25	1	5.25	0.1513	0.6996
A2	2789.47	1	2789.47	80.40	<0.0001
B2	1026.26	1	1026.26	29.58	<0.0001
C2	42.98	1	42.98	1.24	0.2733
Residual	1214.40	35	34.70
Lack of fit	976.17	3	325.39	43.71	<0.0001	Significant
Pure error	238.22	32	7.44
Cor total	5389.64	44

), was statistically significant, with an F-value of 13.37 and a p-value below 0.0001. This confirms the strong influence of the input variables on the hydrocarbon response. The squared terms for the viscosity (A²) and engine speed (B²) were particularly notable, with corresponding F-values of 80.40 and 29.58, respectively. These results emphasize the nonlinear nature of the model and its responsiveness to changes in these parameters. The interaction between the viscosity and fuel type (AC) was significant (F = 8.62, p = 0.0058), demonstrating a strong correlation with HC emissions. Higher-viscosity oils tended to increase hydrocarbon formation, likely due to their impact on fuel atomization and spray dynamics. When the viscosity increased, larger fuel droplets were formed, leading to inefficient mixing with air and incomplete combustion, resulting in higher HC emissions. Conversely, lower-viscosity oils enhanced fuel atomization, promoting a more homogeneous air–fuel mixture, reducing HC formation. In contrast, the individual effects of the viscosity (A) and rpm (B) were not significant (p = 0.1803 and p = 0.2722, respectively), suggesting that their direct influence on HC emissions is lower when considered in isolation. Furthermore, the interaction between the rpm and fuel type (BC) was also insignificant (p = 0.6996). In contrast, the quadratic term for the fuel type (C²) showed a trend toward significance with a p-value of 0.2733 without meeting the required threshold. The model exhibited a significant lack of fit (F = 43.71, p < 0.0001), indicating that further refinement may be necessary to better represent the variability observed in HC emissions. Despite this, the Adeq Precision value of 10.21 reflects a strong signal-to-noise ratio, which supports the model’s suitability for exploring the experimental design space. The coefficient of determination (R²) reached 0.7747, meaning that the model explained 77.47% of the total variation in HC emissions. After adjusting for the number of predictors, the adjusted R² was 0.7167. The predicted R² value of 0.6237 also shows acceptable alignment with the adjusted R², confirming the model’s predictive reliability (refer to

Table 11. Coefficient of determination for HCs.

Coefficient of Determination	Value
R2	0.7747
Adjusted R2	0.7167
Predicted R2	0.6237
Adeq Precision	10.2101

).

Equation (5) presents the final equation using actual factor values for estimating HC emissions. This expression enables the prediction of HC levels based on the specific values of each factor, provided that the input levels are expressed in their original units.

HC [ppm] = −163.9 + 23.2 × Viscosity + 0.04 × RPM − 13.78 × Fuel + 0.000053 × Viscosity × RPM + 1.12 × Viscosity × Fuel − 0.00075 × RPM × Fuel − 0.84 × Viscosity² − 0.000013 × RPM² + 1.97 × Fuel²

(5)

Figure 5a presents the contour plot for HC emissions as a function of the viscosity and rpm. The highest HC values, around 9 ppm, are observed in the central region of the plot, corresponding to intermediate levels of both the viscosity and rpm. This area, highlighted in yellow, represents the operating conditions under which HC formation was most pronounced. As we move towards the plot’s edges, dominated by green and blue tones, a reduction in the HC emissions is observed, reaching values close to 3 ppm.

The response surface (Figure 5b) validates this trend, highlighting a nonlinear dependence of the HC emissions on the viscosity and rpm. Peak HC formation (~9 ppm) was observed at a moderate viscosity (9 mm²/s) and engine speed (1500 rpm), where incomplete combustion was more pronounced. However, at low viscosities (9 mm²/s) or high viscosities (18 mm²/s), HC emissions decreased (~3 ppm). This behavior suggests that combustion efficiency was improved at extreme viscosities through enhanced spray breakup (low viscosity) or stabilized boundary layer formation (high viscosity), leading to better fuel–air mixing and reduced unburned hydrocarbons. This pattern suggests that operating at intermediate values of the viscosity and rpm may promote HC formation, while operation at the upper and lower extremes of these variables could help minimize emissions, possibly due to greater combustion efficiency.

3.3. Model for NOx

The quadratic ANOVA model for NOx emissions demonstrated strong statistical significance, as reflected by an F-value of 38.46 and a p-value below 0.0001, confirming that the selected variables significantly influenced NOx formation (

Table 12. ANOVA results for NOx.

Source	Sum of Squares	Df	Mean Square	F-Value	p-Value
Model	0.0315	9	0.0035	38.46	<0.0001	Significant
A—Viscosity	0.0002	1	0.0002	2.04	0.1623
B—RPM	0.0225	1	0.0225	247.02	<0.0001
C—Fuel	0.0000	1	0.0000	0.4503	0.5066
AB	4.12 × 10−6	1	4.12 × 10−6	0.0453	0.8327
AC	7.83 × 10−6	1	7.83 × 10−6	0.0862	0.7708
BC	0.0035	1	0.0035	38.89	<0.0001
A2	0.0040	1	0.0040	43.50	<0.0001
B2	0.0012	1	0.0012	13.56	0.0008
C2	0.0001	1	0.0001	1.07	0.3077
Residual	0.0032	35	0.0001
Lack of Fit	0.0004	3	0.0001	1.49	0.2349	Not significant
Pure Error	0.0028	32	0.0001
Cor Total	0.0346	44

). Among them, the engine speed (factor B) was identified as the dominant variable, with an F-value of 247.02 and p < 0.0001, underscoring its primary role in generating nitrogen oxides. Moreover, a statistically significant interaction between the engine speed and fuel type (BC) was observed, with an F-value of 38.89 and p < 0.0001, indicating a combined effect on NOx emissions. This suggests that combustion dynamics strongly influence NOx formation, where a higher rpm increases the in-cylinder temperatures, leading to enhanced thermal NOx production. Furthermore, higher-octane fuels can prolong the ignition delay, modify the peak combustion temperatures, and affect NOx emissions. This interaction highlights the need to optimize both parameters to mitigate NOx formation in internal combustion engines.

The quadratic terms of the viscosity (A²) and engine speed (B²) were also significant, with F-values of 43.50 and 13.56, respectively, confirming a nonlinear relationship between these parameters and NOx emissions. This behavior suggests that an increase in the viscosity initially enhances NOx formation, likely due to greater hydrodynamic friction, which increases the engine load and in-cylinder temperatures. However, fuel atomization and the air–fuel mixing efficiency may be affected beyond a certain viscosity threshold, reducing NOx formation. This nonlinearity underscores the complex role of lubrication properties in determining combustion efficiency and emissions control. Conversely, the viscosity (A) and fuel type (C), as well as their individual interactions with other parameters, were not statistically significant (p > 0.1), indicating a limited direct effect on NOx emissions when considered in isolation.

The NOx emission model demonstrated a high explanatory capacity with a determination coefficient (R²) of 0.9082, indicating that it accounted for 90.82% of the observed variability (

Table 13. Coefficient of determination for NOx.

Coefficient of Determination	Value
R2	0.9082
Adjusted R2	0.8846
Predicted R2	0.8584
Adeq Precision	21.3603

). The adjusted R² of 0.8846 compensated for the number of model terms, and the predicted R² of 0.8584 showed strong consistency, validating the model’s predictive robustness. Additionally, the Adeq Precision value of 21.36 confirmed a favorable signal-to-noise ratio, reinforcing the model’s reliability in exploring the design space. The lack-of-fit test produced an F-value of 1.49 with a p-value of 0.2349, indicating that the model’s residual variation was not statistically significant compared to a pure error. This supports the adequacy of the model without the need for further complexity.

The final equation using actual factor values for NOx estimation is presented below as Equation (6):

Nox [% Vol] = 0.95 + 0.03 × Viscosity − 8 × 10⁻⁵ × RPM + 0.033 × Fuel − 1.50 × 10⁻⁷ × Viscosity × RPM − 0.00018 × Viscosity × Fuel − 1.95 × 10⁻⁵ × RPM × Fuel − 0.0010 × Viscosity² + 1.41 × 10⁻⁸ × RPM² − 0.003 × Fuel²

(6)

Figure 6a presents the contour plot for NOx emissions as a function of the viscosity and engine speed. The distribution of NOx concentrations exhibits a transparent gradient, with the highest emissions, approximately 1.10% vol, located in the lower region of the plot, corresponding to high viscosity values (around 18 mm²/s) and increased engine speeds. Conversely, the lowest NOx emissions, approximately 1.04 ppm, appear in the figure’s upper-left and upper-right corners, where the viscosity is at its lowest (around 9 mm²/s) and the engine speed is reduced. This trend suggests that NOx emissions increased with the viscosity and rpm, reaching a peak in an intermediate viscosity range and subsequently stabilizing.

Figure 6b, which shows the response surface plot, confirms the nonlinear relationship between the NOx emissions, viscosity, and engine speed. This plot shows a distinct curvature, confirming that NOx emissions peaked at intermediate-to-high viscosity values and elevated engine speeds. This suggests that an increased oil film thickness at moderate viscosities may contribute to additional frictional heating, raising the in-cylinder temperatures and promoting NOx formation. However, fuel atomization might be hindered at excessive viscosities, altering combustion characteristics and slightly reducing NOx production. This behavior illustrates the intricate relationship between lubrication properties and combustion dynamics in emission formation.

These results suggest that controlling the viscosity and RPM could be crucial for mitigating NOx emissions, as excessive increases in these parameters contributed to more significant pollutant formation. This aligns with the findings from the ANOVA analysis, where the engine speed was identified as the most influential factor in NOx generation. The Response Surface Methodology thus provides an effective tool for determining optimal operating conditions that minimize NOx emissions while maintaining engine efficiency.

3.4. Analysis of Effect of Oil Viscosity on Engine Performance

The viscosity of engine oil plays a key role in determining performance, as it influences the magnitude of mechanical losses associated with fluid resistance within the lubrication circuit, which in turn impacts the power output. To investigate this relationship, dynamometer tests were carried out using three oil grades (SAE 5W-30, 10W-30, and 20W-50) combined with three gasoline types differing in their octane rating (87, 92, and 95). The results obtained from these experiments are detailed in

Table 14. Maximum power measured on the dynamometer for different oils and fuel octane ratings.

Oil	RON	Power [kW]
5W30	87	59.5
	92	58.8
	95	60.2
10W30	87	57.9
	92	56.3
	95	59.5
20W50	87	55.6
	92	56.5
	95	55.9

.

An analysis of the values shown in Table 13 shows that the power output decreased as the oil viscosity increased. The averages for 5W30, 10W30, and 20W50 were 59.50, 57.90, and 56.00 kW, respectively, with less than one-unit standard deviations, ensuring minimal variability in the experiments. The lowest values were recorded for the 20W-50 oil in all cases, indicating a higher mechanical loss in the lubrication system. In contrast, the 5W30 oil provided the highest power in all octane conditions, suggesting an energy reduction in mechanical losses due to internal friction.

The trends observed in Table 14 are corroborated by the power curves shown in Figure 7, which present the power profiles as a function of the engine speed for each oil viscosity and fuel type.

An analysis of the curves reveals that increasing the octane rating of the fuel resulted in a slight increase in the engine power. In particular, 95-octane fuel showed an average 1.49% benefit compared to 87-octane fuel for all samples, which can be attributed to more efficient combustion and the engine’s ability to operate at higher compression ratios without the risk of premature detonation. However, the influence of the octane rating was less significant than that of the oil viscosity, which remained the dominant factor in determining the engine power.

These results align with those from the pollutant emissions analysis, which identified that using lower-viscosity oils improved the engine power and reduced CO₂ and NOx emissions. The decrease in the mechanical loss due to internal friction allowed for improved combustion, reducing the amount of unburned residues and optimizing the overall engine performance.

In conclusion, the selection of lubricating oil significantly impacts the engine power. Its selection should not only consider the protection and durability of the lubrication system but also the energy efficiency and emissions generated. Combining a low-viscosity oil with a higher-octane fuel represents an optimal strategy to improve engine efficiency without compromising engine performance.

4. Discussion

The results of this study confirm that the oil viscosity and fuel quality significantly impact engine performance and pollutant emissions. In terms of the engine output, the data obtained from the test bench revealed that using a low-viscosity oil (SAE 5W-30) resulted in a higher engine power, with an average output of 59.50 kW. In contrast, the higher-viscosity oil (SAE 20W-50) yielded the lowest power output (56.00 kW). This suggests that high-viscosity oils induce more significant mechanical friction losses, reducing the overall engine efficiency. Regarding fuel quality, using 95-octane gasoline led to an average power increase of 1.49% compared to using 87-octane gasoline, which can be attributed to improved combustion efficiency and a lower tendency to knock.

The emissions analysis demonstrated a direct correlation between the input variables and the CO₂, NOx, and HC emissions. Notably, the highest CO₂ levels (~14.4% vol) were observed when using high-viscosity oils (18.5 mm²/s) and operating at high RPMs (2500 rpm), suggesting increased fuel consumption due to higher internal engine friction. In contrast, using low-viscosity oil (9.66 mm²/s) at low RPMs (750 rpm) reduced CO₂ emissions to 13.4% vol, indicating lower hydrodynamic resistance in lubrication and enhanced thermal efficiency. However, it is important to note that higher CO₂ emissions do not necessarily imply inefficiency. A more complete combustion process typically results in higher CO₂ concentrations, as CO₂ is the primary oxidation product of carbon-based fuels under stoichiometric conditions. Previous studies have shown that fuel formulations with a higher aromatic content can increase the THC and CO₂ emissions due to modified combustion characteristics, particularly in gasoline direct injection engines. Thus, while in this study, elevated CO₂ emissions suggested higher fuel consumption under certain conditions, they may also indicate improved combustion efficiency, depending on the operating parameters and fuel properties.

The statistical analysis, based on an ANOVA, confirmed that the engine speed (RPM) was the most influential factor in emissions, with F-values of 163.11 (p < 0.0001) for CO₂ and 247.02 (p < 0.0001) for NOx, indicating a highly significant effect. The interaction between fuel quality and the RPM was also relevant for CO₂ formation (F = 15.11, p = 0.0004), highlighting that higher-octane fuels may contribute to improved combustion efficiency at high speeds.

The findings of this study align with and extend previous research on the influence of the oil viscosity and fuel quality on engine performance and emissions. Similarly to the results reported by Zhang et al. [9], the use of low-viscosity oil (SAE 5W-30) contributed to improved engine power, supporting the concept that reduced hydrodynamic friction enhances mechanical efficiency. In line with Fernanda et al. [2] and Rodríguez-Fernández et al. [1], higher-octane fuels led to a modest but consistent increase in the power output, attributed to more stable combustion and a resistance to knock. Regarding emissions, the observed increase in NOx levels at higher RPMs is consistent with the findings of Iodice et al. [7], who associated elevated combustion temperatures with higher NOx formation. However, the lack of statistical significance for the role of the oil viscosity in the CO₂ model contrasts with studies such as that by Giakoumis [13] which report a more direct correlation between lubricant properties and fuel economy. This divergence may be due to the relatively narrow viscosity range tested or the specific operating conditions used in this study. Overall, the integration of real-world fuels and controlled dynamometer testing contributes practical value to these observations, particularly in the context of regional markets with variable fuel standards.

These results can be explained by underlying thermodynamic and tribological principles. The higher power output associated with low-viscosity oils results from reduced internal friction, which lowers mechanical losses in the engine. Similarly, increased NOx emissions with a higher engine speed can be attributed to higher in-cylinder temperatures, which enhance thermal NOx formation. The improved performance with higher-octane fuels stems from their greater resistance to knock, allowing for more efficient combustion under advanced ignition timing.

These results affect the automotive industry, lubricants manufacturing, and fuel formulation. The data suggest that combining low-viscosity oils with higher-octane fuels can improve engine efficiency and reduce emissions, contributing to the development of more sustainable technologies. However, it is crucial to consider the engine type and operating conditions, as the engine’s response may vary depending on its design and combustion system.

Future research could evaluate a broader range of lubricants, including ultra-low-viscosity oils (SAE 0W-20), and analyze the impact of alternative fuels, such as ethanol blends in varying proportions or synthetic biofuels. Additionally, long-term studies on engine wear should be conducted, considering the effect of different oil formulations on the durability of mechanical components. Ultimately, examining the impact of the oil viscosity on hybrid vehicles would offer valuable insights into lubrication performance under intermittent operating conditions.

5. Conclusions

This study analyzed the combined effects of the oil viscosity, fuel octane rating, and engine speed on an internal combustion engine’s performance and pollutant emissions under controlled dynamometer conditions. The results demonstrate that optimizing lubrication and fuel parameters can improve engine efficiency and reduce emissions. Among the studied factors, the engine speed greatly influenced NOx and CO₂ emissions, while lower-viscosity oils were associated with a higher power output and reduced frictional losses. However, the experimental design did not include engine load or direct fuel consumption measurements, which limits the scope of interpretation regarding the real-world efficiency. These aspects are essential to include in future studies that will offer a more complete understanding of the energy–emissions trade-off. It is recommended that future research explore a broader range of lubricant formulations, including ultra-low-viscosity oils, and evaluate alternative fuels such as ethanol blends and synthetic biofuels. Furthermore, assessing lubrication performance under variable load conditions and in hybrid vehicles may provide valuable insights for sustainable powertrain development.