An Evaluation of Predictive Modeling and Error Analysis for

Physicochemical Characteristics of Gasoline-Bioethanol Blends in

a Gasoline Engine
1
Ogunsola A.D. 2Alajede A. M, 3Olafimihan E.O, 4Sangotayo E.O*,
5
Aderibigbe A. A. & 6Sulaiman A. O.
1,2,3,4,5,
Department of Mechanical Engineering, Ladoke Akintola University of Technology, Ogbomoso, Nigeria,
6
Department of Aeronautics and Astronautics, Kwara State University, Malete, Kwara, Nigeria.,
*
Corresponding author: [email protected]

ABSTRACT

The integration of bioethanol into gasoline blends has gained significant attention for improving engine
performance and reducing environmental impacts. This study evaluated predictive modeling and error analysis
of the physicochemical characteristics of gasoline-bioethanol blends (E0, E5, E10, and E15) in a four-stroke
gasoline engine. The importance of optimizing blend ratios for efficiency and sustainability underscores the
relevance of this research. The physicochemical characteristics, including cetane number, viscosity, density,
carbon residue, and heating values, were measured using a data logger and analyzed across blend ratios ranging
from 0% to 15%. Profiles of these characteristics were created to assess their relationship with bioethanol
content. Predictive models were developed using SAS software, with R-squared and RMSE values evaluated as
performance metrics to assess model accuracy and fitness.

The results indicate that the coefficient of performance (COP) demonstrated higher sensitivity to bioethanol
content in the 0-10% range, stabilizing beyond 10%, suggesting an optimal blend ratio of around 10%. Brake
power efficiency decreased linearly with increasing bioethanol content due to the lower energy density of higher
ethanol blends. However, the enhanced combustion efficiency of bioethanol compensated for some efficiency
losses. Notably, the E10 blend achieved the highest brake power of 391.65 W, Model evaluation revealed robust
predictive capabilities, with R-squared and RMSE values for COP at 0.964 and 0.585, respectively, and for
braking thermal efficiency at 0.996 and 0.133, respectively. These metrics confirm the model's high accuracy
and reliability in predicting engine performance characteristics across blend ratios. Future research should
explore the impact of higher ethanol concentrations on engine durability and further optimization of bioethanol-
gasoline formulations for enhanced sustainability and efficiency.

Keywords: Predictive Modeling; Engine Performance; Error Analysis, Physicochemical, SAS, Bioethanol
Blends

INTRODUCTION

The increasing need for sustainable and eco-friendly energy sources has heightened the emphasis on renewable
fuels such as bioethanol. Bioethanol, derived from biomass like corn, sugarcane, and cellulosic materials,
provides a renewable substitute for fossil fuels. Combining bioethanol with gasoline diminishes greenhouse gas
emissions and improves fuel characteristics, including octane rating, so enhancing engine performance.
Gasoline-bioethanol mixtures, typically designated as E10 (10% ethanol) or E15 (15% ethanol), are already
utilized worldwide, demonstrating the pragmatic potential of bioethanol in tackling energy and environmental
issues [1]. The incorporation of ethanol, however, dramatically modifies the physicochemical characteristics of
gasoline, encompassing density, viscosity, energy content, and vapor pressure. These modifications affect
combustion properties, engine efficiency, and emissions. Comprehending and forecasting these alterations is
essential for optimizing blend formulations to attain a balance among efficiency, performance, and
environmental sustainability. Predictive modeling provides a robust method for comprehending the intricate
relationships between fuel characteristics and engine performance. Utilizing experimental data and sophisticated
1
computational methods, predictive models can replicate the impact of ethanol concentration fluctuations on the
physicochemical parameters of the mix. As global demand for renewable energy rises, biodiesel is being utilized
in a growing array of sectors as a clean energy source. It can directly substitute or be blended with petroleum
diesel and is extensively utilized in transportation, agricultural machinery, power generation equipment, and
several other sectors. In recent years, numerous nations and areas have advanced the production and utilization
of biodiesel via policy incentives and regulatory mandates to diminish reliance on fossil fuels and mitigate
greenhouse gas emissions [,2,3].

Biodiesel has emerged as a highly viable alternative fuel among other renewable energy sources due to its
renewability, environmental sustainability, and compatibility with existing diesel engines. Biodiesel is a fuel
derived by the transesterification of vegetable oil, animal fat, or waste oil with alcohols (such as methanol or
ethanol), primarily consisting of fatty acid methyl esters (FAMEs). In comparison to conventional petroleum
diesel, biodiesel not only diminishes greenhouse gas emissions but also significantly lowers the emissions of
carbon monoxide (CO), hydrocarbons (HC), and particle matter (PM), thus alleviating air pollution and
enhancing environmental quality [5]. Globally, the utilization of biodiesel fosters agricultural and industrial
advancement while significantly diminishing reliance on petrochemical fuels [6]. Nonetheless, notwithstanding
the benefits exhibited by biodiesel in the environmental and energy sectors, numerous challenges persist in its
practical implementation. The physicochemical qualities of biodiesel vary considerably based on the raw
materials and production methods, which directly influence its performance in diesel engines [7]. The
combustion characteristics, emission characteristics, and engine performance effects of biodiesel in engines also
fluctuate based on the fuel content. Consequently, precise prediction of biodiesel characteristics and its efficacy
in diesel engines is a primary emphasis of contemporary research [8].

The characteristics of biodiesel directly influence its efficacy in diesel engines. The physicochemical properties
of biodiesel, including viscosity, density, heating value, and oxidation stability, influence not only the
combustion efficiency of the fuel in the engine but also the engine's starting performance, fuel consumption,
emission characteristics, and long-term reliability. The combustion process of biodiesel in diesel engine systems
directly influences thermal efficiency, power production, and emissions. The elevated oxygen concentration in
biodiesel can enhance combustion efficiency and diminish HC and CO emissions [9]. This trait may also result
in elevated nitrogen oxide (NOx) emissions. Consequently, in the optimal design of engines, it is essential to
account for both the physicochemical qualities and combustion characteristics of the fuel. Robust predictive
models must be developed to assess the comprehensive performance of biodiesel. Conventional research
methodologies primarily derive the qualities of biodiesel from experimental measurements; however, this
approach is both time-intensive and expensive, and it is also subject to specific experimental inaccuracies. [10].
With the advancement of computing technology and data science, predictive methodologies utilizing
mathematical models and machine learning (ML) have become significant research domains. These approaches
may swiftly forecast biodiesel characteristics by utilizing current experimental data, offering a solid scientific
foundation for fuel research and application.

The creation of predictive models can significantly reduce trial time and costs while examining the effects of
various raw materials and manufacturing methods on fuel performance, therefore optimizing production and
enhancing fuel quality [11]. Prediction models utilizing linear regression, support vector machines (SVMs), and
artificial neural networks (ANNs) can precisely forecast essential characteristics such as heating value,
viscosity, and oxidation stability of biodiesel [12]. The utilization of these models allows researchers to
comprehend the attributes of various biodiesel fuels, offering efficient instruments for fuel optimization design.
Simultaneously, a growing contingent of researchers is employing data-driven methodologies to forecast the
performance of biodiesel in engines. Through the analysis of extensive experimental data and model training,
machine learning techniques may elucidate the intricate nonlinear correlations between fuel characteristics and
engine performance, resulting in very accurate forecasts. These methodologies offer novel insights into the
research of biodiesel-engine interactions. They provide substantial assistance in optimizing fuel mixtures and
engine design.

2
Biodiesel, as a crucial renewable energy source, has substantial environmental advantages and application
potential due to its renewability, reduced carbon dioxide emissions, and favorable biodegradability. Biodiesel
research has advanced swiftly, with scholars consistently broadening the sources of raw materials, fuel blend
compositions, and application contexts for biodiesel [13,14]. The efficacy of biodiesel fluctuates based on the
raw materials utilized. Moreover, biodiesel is generally combined with diesel through the use of additives for
practical application. Consequently, the characteristics of biodiesel blends and their efficacy in diesel engines
are more intricate than those of other alternative fuels. This article seeks to systematically compile and evaluate
recent scientific advancements in forecasting biodiesel characteristics and its performance in diesel engines.
Beginning with the essential characteristics and performance metrics of biodiesel, particular research instances
were referenced, encompassing application examples of various predictive models and experimental techniques.
This review elucidates the merits and drawbacks of different methodologies in empirical research. These
scenarios facilitate the identification of optimal practices and prospective enhancements in existing
methodologies, offer direction for future study, and encourage the advancement of biodiesel applications. This
situation has initiated the present research, which focuses on the assessment of predictive modeling and error
analysis for physicochemical characteristics of gasoline-bioethanol blends in a gasoline engine

MATERIALS AND METHODS

Materials and Equipment
This study utilized the following materials and components: a 35-liter Bed Side refrigeration system, a 0.5mm
thick aluminum light sheet, automotive paint in Ash color, a 16-gauge electrode, a capillary tube, a hose, fiber,
resin, wheels, bioethanol fuel (quantity and composition), and refrigerant (R134a).

Construction and Development of a Compartment

The bio-ethanol engine, automotive air conditioner, and vaccine storage are the three systems that constitute the
compact refrigeration module. The automotive air conditioner is directly connected to the bio-ethanol generator.
The refrigeration unit that is implemented is a modification of the original components of an automotive vapor-
compression air-conditioning system, which consists of a compressor, condenser, expansion valve, and
evaporator, as well as an oil chiller and oil separator. The movement of air streams facilitates the exchange of
heat between the condenser and evaporator. The thermostatic type of expansion valve is implemented. The
refrigerant for this investigation will be R-134a. The indoor coil, which serves as an evaporator, absorbs heat
from the fan-driven air stream as the refrigerant passes through it, thereby delivering refrigeration. The
refrigerant is introduced to the outdoor coil after compression, where it dissipates heat into an additional air
stream. The automotive air conditioning system is capable of functioning in a refrigeration cycle.
A 4-stroke, single-cylinder gasoline engine with a power output of 5 to 10 horsepower was implemented. This
engine supplied the mechanical energy necessary to operate the refrigeration compressor. Equation (1)
determines the engine's I.C. brake output.

ω is the angular speed in radians per second (rad/sec.)

Thus,

Equation 3 gives the expression for the brake thermal efficiency.

Where PB is the brake power in watts, mf is the mass flow rate of the fuel by this engine in kg/s, CV is the
heating value used by this engine in J/kg.

Refrigeration System
Figure 1 illustrates the refrigeration system, which includes a compressor, refrigerant, heat exchangers, and
condenser. The refrigeration system is depicted in Figure 2 from the orthogonal and isometric perspectives, as
well as from the front, end, and top views.

3
 i. Refrigeration Compressor - a reciprocating compressor that is hermetically sealed and compatible with
the engine's power output was chosen. The refrigerant's pressure and temperature were increased by
compressing it with the compressor.
ii. Refrigerant - R134a, an environmentally benign refrigerant, was used to charge the system. This was
charged by the system's specifications.
iii. Heat Exchangers - These were employed to transfer heat from the refrigerated chamber to the external
environment.
iv. Condenser - A condenser that is air-cooled and has an adequate heat rejection capacity was employed.
The purpose of this was to transfer heat from the refrigerant to the ambient air.
v. Evaporator - A fin-tube-type evaporator was implemented. This functioned to absorb heat from the
refrigerated space.
vi. Thermostatic Expansion Valve (TXV): A Thermostatic Expansion Valve (TXV) was installed. This
regulated the passage of refrigerant into the evaporator.
vii. Refrigeration Chamber: This is an insulated chamber with a volume of approximately 1-2 cubic meters
that was employed. This functioned as the experiment's cooling compartment.

Figure 1: Refrigeration System

Figure 2: Orthogonal and Isometric View of the Refrigeration System

METHODOLOGY FOR SYSTEM ASSEMBLY

During the system's assembly, a suitable coupling or belt drive system was implemented to establish a
connection between the refrigeration compressor and the petroleum engine. Using copper tubing, the
compressor discharge line was connected to the condenser intake. The expansion valve outlet was connected to
the evaporator inlet, and the expansion valve inlet was connected to the condenser outlet. The evaporator output
was reconnected to the compressor vacuum line. To ensure that all connections were secure and leak-proof, they
were examined. Vacuum pumps were implemented to evacuate the system, thereby eradicating any air and
moisture. Subsequently, the refrigerant (R134a) was introduced into the system.

EVALUATION OF PERFORMANCE

4
The compressor vacuum and discharge pressures, temperatures at critical locations, refrigerant flow rate, and
fuel consumption rate of the gasoline engine were monitored and recorded. The temperature and flow rate data
that were obtained were analyzed to ascertain the cooling capacity of the system.
Coefficient of Performance (COP) was computed using Equation 4.

The Gasoline-driven refrigeration system's efficacy was assessed through data analysis.

Data Collection
The data logger was utilized to collect data on temperature, pressure, and flow rate throughout the testing
period. The cooling capacity, coefficient of performance (COP), and fuel efficiency were computed. Theoretical
predictions were compared with experimental outcomes. The data was examined to discern trends, including the
impact of engine strain on system performance.

Technique of model assessment

The correctness of the correlations was evaluated using two widely employed statistics metrics: Root
Mean Square Error (RMSE) and Mean Absolute Bias Error (MABE). (MJ/m 2) for the regression model as
presented in Eq.(5-7), [16,16,17]

(5)

(6)

(7)

The RMSE test offers insight into the short-term efficacy of correlations by facilitating a term-by-term
comparison of the actual discrepancies between estimated and measured values; a smaller RMSE indicates a
more precise estimate. A positive MBE indicates overestimation, whereas a negative number denotes
underestimated by the model. A disadvantage of this approach is that an overestimation of one observation
negates an underestimation of another, resulting in MBE values that demonstrate systematic error or bias, while
RMSE reflects a non-systematic error [18]. The research assessed the correlations and regressions among the
monthly mean diffuse solar radiation, clearness index, relative sunshine duration, circulation index, and the ratio
of minimum to maximum temperature. The coefficient of determination (R²) seeks to attain a value of 1, ideally
reaching 100% to ensure greater accuracy and reliability in data modeling [19].

RESULT AND DISCUSSION

Micro Cooling System Assembly Developed
Micro-cooling systems are designed to maintain the appropriate temperature for preservation. Figure 3 shows
how the developed micro-cooling system was assembled. The assembly includes the insulated compartment,
cooling unit, temperature control system, data logging and monitoring, evaporator outlet and blower motor with
protective cover, condenser with fan, dryer, alternator, freezer compartment, base and base tire, vaccine tray,
and working substance.
Micro-cooling systems reduce heat transfer from the outside via an enclosed compartment. The highest-quality
thermal insulation materials utilized in this compartment maintain the internal temperature regardless of ambient
conditions. The compartment's cooling technology was designed to keep the air temperature within the vaccine
storage range. The energy-efficient cooling unit runs for a long time without using much power. A precise
temperature control system with digital thermostats and sensors was also developed. These components
continuously monitor the interior temperature to modify cooling output. This ensures that temperature
fluctuations from the ideal range (2°C to 8°C) are quickly corrected, maintaining reliable preservation
conditions. The system logs data for safety and compliance. This tool tracks temperature changes to monitor
vaccine storage.
The micro refrigeration system can run on AC mains and battery backup using a gasoline-bioethanol blend.
Immunizations can be preserved in locations with intermittent energy, making them suitable for distant areas or
emergencies. A crucial healthcare invention, the gasoline-driven vaccine refrigeration system (Figure 3), ensures
5
appropriate vaccine storage and transportation. Its strong architecture and precise temperature management help
preserve vaccine efficacy and public health and safety.

Figure 3: Assemblage of Gasoline-driven Refrigeration System

Physicochemical Properties of Gasoline-Bioethanol Blends

Figures 4 to 12 show that each physicochemical feature fluctuated as the volume fraction of ethanol in the
combinations rose. The gasoline-bioethanol mixture increased density, viscosity, heating parameters, and
specific gravity. Predictive models of the properties are shown on the plot and error analysis of models are
evaluated as function MSE and RMSE.
Figure 4 shows octane numbers against blending ratios. Fuel mixtures with ethanol blends from 6% to 15% have
higher octane ratings than those with 0 to 6%. At a 6% blend ratio, the minimum octane is 8°C. This helps
reduce ringing in high-compression engines, which need higher-octane fuels. Ethanol burns cooler and
consumes slower than gasoline, allowing for more regulated combustion. This improves the blend's effective
octane rating, reducing pre-ignition and clanging.

Figure 4: Trend of Octane Number against Blending Ratios

Figure 5 shows the link between blending ratios and centipoise viscosity at 20 °C. Viscosity rises from 0 to 6%,
then drops as ethanol is added to gasoline, raising the blend ratio from 6% to 15%. At a 6% mix ratio, viscosity
peaks at 14 centipoise. Additionally, ethanol is more viscous than petroleum. Blending increases fuel mixture
viscosity (Figure 5). This change may affect engine combustion and fuel atomization. In cooler temperatures,

6
ethanol may have higher fuel viscosity than pure gasoline, reducing engine performance and fuel delivery. At
operational temperatures, viscosity may stabilize, but the compound will still be thicker than gasoline.

Figure 5: Trend of Viscosity centipoise @ 20 °C against Blend Ratio (%)

In Figure 6, Flashpoint is graphed against blending ratios. Ethanol blends with gasoline from 6% to 15%
increase the fuel mixture's flashpoint. At 6% mix ratio, 25°C is the minimal flashpoint. Blended gasoline is less
volatile than pure gasoline because its flash point rises. This can improve storage and handling safety. The
addition of ethanol to gasoline should change its volatility. Ethanol reduces gasoline's lighter components'
volatility, raising the blend's flash point. This change may lessen vapor lock and improve high-temperature
safety.

Figure 6: Trend of Flashpoint versus Blending Ratios

Density-blending ratio correlation is shown in Figure 7. Density increases from 0 to 6%, then declines when
ethanol is combined with gasoline, boosting the blend ratio from 6% to 15%. At 6%, the density reaches 0.89
kg/L. Ethanol reduces gasoline density (Figure 7). Engine fuel supply and combustion may be affected by
density reduction. The fuel density reduces as ethanol content increases. This may vary fuel volumetric flow
rate, influencing engine performance and efficiency. The combination's density increased.

7
 Figure 7: Trend of Density (kg/L) versus Blend Ratio (%)

Figure 8 shows the carbon residue-blending ratio graph. The carbon residue in the fuel mixture reduces from 0%
to 10% but increases when 10% to 15% ethanol is added to gasoline. The minimum carbon residue is 74% at a
10% mixing ratio. Ethanol has less carbon and more oxygen than gasoline, which reduces carbon residue at
lower blend ratios, but gasoline's high carbon content persists at lower ethanol concentrations. The blend with
10% ethanol has the lowest Carbon residue (74%), indicating excellent mixing. However, adding more
complicated compounds may increase Carbon residue. Lower carbon residue reduces particulate matter and
CO2 emissions, while optimum blending ratios boost engine efficiency and reduce maintenance.

Figure 8: Trend of Carbon residue versus Blend Ratio (%)

Blending ratios and Specific Gravity (SG) are shown in Figure 9. Specific gravity rises from 0 to 6%, then drops
when ethanol is added to gasoline, raising the blend ratio from 6% to 15%. At 6% blend ratio, specific gravity
peaks at 0.85. Specific gravity affects fuel density, viscosity, and combustion properties. Ideal blending ratios
increase engine efficiency, power output, and emissions, and understanding SG-blend ratio connections helps
refine fuel specifications. Suboptimal interactions between ethanol and gasoline at high concentrations, the
surplus of ethanol increasing volume and decreasing density, and molecular arrangements that affect mixture
density may reduce specific gravity beyond the 6% blend ratio.

8
 Figure 9: Trend of Specific Gravity @ 20oC versus Blend Ratio (%)

Figure 10 shows how mixing ratios affect pour point oC. As the blend ratio rises from 0% to 15%, the fuel
mixture's pour point oC fluctuates. Pour point oC peaks at 8oC at a 10% blend ratio. It shows that ethanol's
lower viscosity and higher volatility affect the blend's Pour Point. At 0-5% ethanol, gasoline dominates the
combination. Ethanol and gasoline molecules interact to raise the Pour Point (5-10%). The blend's Pour Point
(8°C) at 10% ethanol indicates optimal mixing. Over 10% pour point decreases: The Pour Point may be
decreased by adding 10-15% ethanol, which may disrupt molecular connections. Pour Point may affect gasoline
flowability, especially in cooler conditions. Optimized blending ratios ensure engine reliability, while Pour
Point-blend ratio relationships guide fuel storage and management.

Figure 10: Trend of Pour point oC versus Blend Ratio (%)

The cetane number and blending ratios are shown in Figure 11. The Cetane number of the fuel combination
varies from 0% to 15%, peaking at 15°C at a 10% blend ratio. Higher octane and lower cetane of ethanol affect
blend cetane. At ethanol concentrations of 0–5%, the combination behaves like gasoline; at 5–10%, ethanol and
gasoline molecules interact, raising the Cetane Number. At 10% ethanol, the mix has the maximum Cetane
Number, indicating optimal amalgamation. A higher Cetane Number improves gasoline ignition and reduces
engine knocking, and optimum blending ratios reduce emissions and efficiently burn fuel.

9
 Figure 11: Trend of Cetane number oC versus Blend Ratio (%)

Figure 12 shows the blending ratio-heating values graph. Fuel combination heating values increase from 0 to
10% but decrease from 10% to 15% ethanol to gasoline. The maximal heating value is 410 MJ/kg at a 10%
blend ratio. Figure 12 shows the relationship between gasoline-bioethanol compounds' higher and lower thermal
values. The highest heating value (410-412 kJ/kg) indicates the energy content of the E10 gasoline-bioethanol
combination. This is like testing fuel-efficient and power-producing mixes. As gasoline contains E15
bioethanol, the composite's thermal value decreases. The blend's low energy value may explain this. Reduced
heating value may affect combustion efficiency. Lower-energy mixture engines may need timing and fuel
injection changes to function better. If calibrated properly, this can complete combustion and boost efficiency.
The diluting impact of lower-energy ethanol in gasoline lowers the blend's thermal value (Figure 12). Heating
value decreases may also reduce combustion efficiency. The lower-energy combination may require fuel
injection and timing changes to improve engine performance. If tuned properly, this can complete combustion
and boost efficiency.

Figure 12: Trend of Heating Values versus Blend Ratio (%)

Characterization of Engine Test ResultsforE0, E5, E10 and E15 Blends

Speed, torque, braking power, brake thermal efficiency, and COP from the four-stroke spark-ignition engine test
rig. Results are shown in Figures 13-15. The finding was used to determine the engine test rig's gasoline-
bioethanol performance parameters. Figure 13 shows torque (Nm), Figure 14 shows braking power (kW), and
Figure 15 shows brake thermal efficiency (%) and COP from the engine test rig running on a gasoline-
bioethanol blend at 356.19 rad/sec.

10
Analyzing engine performance data at 356.19 rps while adjusting torque, braking power, and brake thermal
efficiency.

The E10 gasoline-bioethanol blend produces 10.8 Nm, whereas the E15 produces 9.8 Nm. This means the
engine can create more rotational force in the third condition, improving acceleration and performance. The E0
gasoline-bioethanol blend has 373.00 W of brake power, while the E5 blend has 384.19 W, the E10 blend
391.65 W, and the E15 blend 365.54 W.

Figure 13 shows that torque (Nm) declines with the gasoline-bioethanol ratio. Bioethanol has a lower energy
density (~26.7 MJ/kg) than gasoline (~44.4 MJ/kg), and the presence of oxygen may restrict energy release
during combustion.
Lower torque reduces engine power and performance, but lower energy density and changed combustion
characteristics may reduce fuel efficiency and impair acceleration and responsiveness.

Figure 13: Trend of Torque (Nm) versus Blend Ratio (%)

Figure 14 shows how the gasoline-to-bioethanol blend ratio affects brake power (kW). Brake Power (kW)
decreases non-linearly with the bioethanol mix ratio, with a greater loss at higher bioethanol concentrations. The
study shows that bioethanol's lower energy density (~26.7 MJ/kg) compared to gasoline (~44.4 MJ/kg) reduces
Brake Power. Bioethanol's higher octane rating and combustion properties may minimize combustion energy.
Bioethanol's lower energy content reduces brake power. Bioethanol's high octane rating can cause engine knock
or pinging, reducing Brake Power and fuel efficiency. Low Brake Power reduces vehicle acceleration,
responsiveness, and performance, lowers fuel efficiency, especially at high bioethanol mix ratios, and damages
engine components over time.

Figure 14: Trend of Brake power (kW) versus Blend Ratio (%)
11
Figure 15 shows that the gasoline-to-bioethanol blend ratio linearly decreases braking power efficiency. The
coefficient of performance (COP) declines from 0% to 10% as the gasoline-to-bioethanol blend ratio increases,
but 12% remains stable beyond 10%. Braking power efficiency decreases linearly with the bioethanol mix ratio,
therefore each percentage point increase in bioethanol reduces this efficiency. As the bioethanol blend ratio
increases from 0% to 10%, the coefficient of performance (COP) decreases, but it stabilizes at 6% above 10%.
Bioethanol requires engines to adjust to their combustion properties at high mix percentages and blends above
10% increase performance and reduce losses. High bioethanol concentrations increased combustion efficiency,
compensating for energy density decreases, although the long-term effects on engine components need to be
assessed. The brake power of the E10 gasoline-bioethanol blend peaks at 391.65 W, which is the engine's output
power after friction and other losses. The E10 gasoline-bioethanol combination (8.16%) has higher brake power,
indicating a more efficient engine at the same speed.

Figure 15: Trend of braking thermal efficiency and Coefficient of Performance (COP) of the system versus
Blend Ratio (%)

Table 1.0 presents a summary of the predictive model with R-squared and RSME. RMSE quantifies the
disparity between expected and actual values. It computes the square root of the mean of the squared deviations
between expected and actual values. Low RMSE values signify that the model effectively fits the data and
yields more precise forecasts, whereas high RMSE values denote greater inaccuracy and reduced forecasting
precision. RMSE values ranging from 0.2 to 0.5 indicate that the model can predict data with acceptable
accuracy, whereas lower values signify superior model fit and enhanced prediction accuracy. Adjusted R-
squared quantifies the proportion of variance in the dependent variable elucidated by the independent variables,
with elevated Adjusted R-squared values signifying a high degree of precision. Criteria for Modified R-squared
Values A value of 0.75 or above signifies a high degree of accuracy, while a value of 0.4 or higher is
considered acceptable in some circumstances. RMSE and Adjusted R-squared serve as complementing metrics
for assessing model performance. RMSE quantifies the discrepancy between predicted and actual values,
whereas Adjusted R-squared assesses the fraction of variance elucidated by the model.

Table 1.0 Summary of the predictive model with R-squared and RSME

Properties Model for Blend Ratio (0%-15%) Coefficient of RMSE

Determination
Octane Number y = -0.012x3 + 0.36x2 - 2.683x + 14 R² = 1 1.176982

Viscosity y = 0.003x3 - 0.082x2 + 0.531x + 0.411 R² = 1 0.27155743

Flashpoint y = -0.018x3 + 0.54x2 - 3.733x + 32 R² = 1 1.176982

12
 Density y = 1E-04x3 - 0.003x2 + 0.024x + 0.835 R² = 1 0.041391

Carbon Residue % y = 0.13x2 - 2.13x + 83.35 R² = 0.949 0.782624

Specific Gravity y = -0.002x2 + 0.016x + 0.829 R² = 0.995 0.076575

Pour point y = -0.018x3 + 0.42x2 - 2.233x + 7 R² = 1 1.176982

Cetane number y = -0.018x3 + 0.44x2 - 2.333x + 13 R² = 1 1.176982

Torque y = -0.076x + 4.87 R² = 0.925 0.533772

Brake power y = -0.000x3 + 0.004x2 - 0.038x + 1.658 R² = 1 0.384199

Braking thermal y = -0.416x + 22.58 R² = 0.996 0.132853

efficiency
Coefficient of y = 0.058x2 - 1.353x + 18.58 R² = 0.964 0.585587
Performance (COP)

Low RMSE values indicate that the model effectively fits the data and generates more precise predictions.
Conversely, increased values suggest a decrease in forecast accuracy and a greater degree of inaccuracy. The R-
squared is a standardized equivalent of the root mean square error, which is a non-standardized measure of
goodness-of-fit. The majority of the models in Table 1.0 have RMSE values ranging from 0.2 to 0.5, which
indicates that the model can predict the data with reasonable accuracy. Additionally, an R-squared value of 0.75
or higher indicates a high level of accuracy, according to a general guideline. The RMSE and R-squared values
of the Coefficient of Performance (COP) are 0.585587 and 0.964, respectively. The RMSE and R-squared of the
braking thermal efficiency are 0.132853 and 0.996, respectively.

CONCLUSIONS
The following conclusions were derived from the present investigation on an evaluation of predictive modeling
and error analysis for physicochemical characteristics of gasoline-bioethanol blends in a gasoline engine. The
results revealed that bioethanol significantly influences the engine's performance metrics. The coefficient of
performance (COP) displayed high sensitivity to bioethanol content in the 0-10% range and stabilized beyond
10%, indicating an optimal blend ratio around E10. While brake power efficiency decreased linearly with
increasing bioethanol content due to its lower energy density, enhanced combustion efficiency at higher ethanol
ratios mitigated these losses. The E10 blend emerged as the most practical and efficient option, achieving the
highest brake power of 391.65 W.

The predictive models developed in this study demonstrated strong accuracy, with R-squared values of 0.964 for
COP and 0.996 for braking thermal efficiency, alongside RMSE values of 0.585 and 0.133, respectively. These
metrics validate the reliability of the models in predicting engine performance across varying blend ratios. The
findings emphasize the E10 blend as an optimal solution for balancing engine performance and environmental
benefits. This study contributes to the growing body of research on bioethanol-gasoline blends and highlights
the potential for their broader application in sustainable energy systems.

Conflicts of Interest

The authors declare no conflicts of interest.

REFERENCES

1. Su, Y.; Zhang, P.; Su, Y. An Overview of Biofuels Policies and Industrialization in the Major Biofuel
Producing Countries. Renew. Sustain. Energy Rev. 2015, 50, 991–1003
2. De Souza, T.A.Z.; Pinto, G.M.; Julio, A.A.V.; Coronado, C.J.R.; Perez-Herrera, R.; Siqueira, B.O.P.S.;
da Costa, R.B.R.; Roberts, J.J.; Palacio, J.C.E. Biodiesel in South American Countries: A Review on

13
 Policies, Stages of Development and Imminent Competition with Hydrotreated Vegetable Oil. Renew.
Sustain. Energy Rev. 2022, 153, 111755.

3. Giakoumis, E.G.; Dimaratos, A.M.; Rakopoulos, C.D.; Rakopoulos, D.C. Combustion Instability
during Starting of Turbocharged Diesel Engine Including Biofuel Effects. J. Energy Eng. 2017, 143,
4016047.

4. Setiawan, I.C.; Setiyo, M. Renewable and Sustainable Green Diesel (D100) for Achieving Net Zero
Emission in Indonesia Transportation Sector. Automot. Exp. 2022, 5, 1–2.

5. Maawa, W.N.; Mamat, R.; Najafi, G.; De Goey, L.P.H. Performance, Combustion, and Emission
Characteristics of a CI Engine Fueled with Emulsified Diesel-Biodiesel Blends at Different Water
Contents. Fuel 2020, 267, 117265

6. Giakoumis, E.G.; Rakopoulos, C.D.; Dimaratos, A.M.; Rakopoulos, D.C. Combustion Noise Radiation
during the Acceleration of a Turbocharged Diesel Engine Operating with Biodiesel or N-Butanol
Diesel Fuel Blends. Proc. Inst. Mech. Eng. Part D J. Automob. Eng. 2012, 226, 971–986

7. Khan, I.U. Analysis of Biodiesel and Fatty Acids Using State-of-the-Art Methods from Non-Edible
Plants Seed Oil; Nicotiana Tobaccum and Olea Ferruginia. Process Saf. Environ. Prot. 2024, 186, 25–
36.

8. Krishnasamy, A.; Bukkarapu, K.R. A Comprehensive Review of Biodiesel Property Prediction Models
for Combustion Modeling Studies. Fuel 2021, 302, 121085

9. Palani, Y.; Devarajan, C.; Manickam, D.; Thanikodi, S. Performance and Emission Characteristics of
Biodiesel-Blend in Diesel Engine: A Review. Environ. Eng. Res. 2022, 27, 200338.

10. Bukkarapu, K.R.; Krishnasamy, A. A Critical Review on Available Models to Predict Engine Fuel
Properties of Biodiesel. Renew. Sustain. Energy Rev. 2022, 155, 111925

11. Suvarna, M.; Jahirul, M.I.; Aaron-Yeap, W.H.; Augustine, C.V.; Umesh, A.; Rasul, M.G.; Günay,
M.E.; Yildirim, R.; Janaun, J. Predicting Biodiesel Properties and Its Optimal Fatty Acid Profile via
Explainable Machine Learning. Renew. Energy 2022, 189, 245–258

12. Xiao, H.; Wang, W.; Bao, H.; Li, F.; Zhou, L. Biodiesel-Diesel Blend Optimized via Leave-One Cross-
Validation Based on Kinematic Viscosity, Calorific Value, and Flash Point. Ind. Crops Prod. 2023,
191, 115914

13. Yusoff, M.N.A.M.; Zulkifli, N.W.M.; Sukiman, N.L.; Chyuan, O.H.; Hassan, M.H.; Hasnul, M.H.;
Zulkifli, M.S.A.; Abbas, M.M.; Zakaria, M.Z. Sustainability of Palm Biodiesel in Transportation: A
Review on Biofuel Standard, Policy and International Collaboration Between Malaysia and Colombia.
Bioenerg. Res. 2021, 14, 43–60.

14. Rakopoulos, C.D.; Antonopoulos, K.A.; Rakopoulos, D.C.; Hountalas, D.T.; Giakoumis, E.G.
Comparative Performance and Emissions Study of a Direct Injection Diesel Engine Using Blends of
Diesel Fuel with Vegetable Oils or Bio-Diesels of Various Origins. Energy Convers. Manag. 2006, 47,
3272–3287.

15. Gebremariam, S.N.; Marchetti, J.M. Economics of Biodiesel Production: Review. Energy Convers.
Manag. 2018, 168, 74–84.

16. Filho, E.P.M., Oliveira, A.P., Vita, W.A., Mesquita, F.L.L., Codato, G., Escobedo, J.F., Cassol, M. and
Franca, J.R.A. (2016) Global, Diffuse and Direct Solar Radiation at the Surface in the City of Rio de
Janeiro: Observational Characterization and Empirical Modeling. Renewable Energy, 91, 64-74
17. Despotovic, M., Nedic, V., Despotovic, D. and Cvetanovic, S. (2016) Evaluation of Empirical Models
for Predicting Monthly Mean Horizontal Diffuse Solar Radiation. Renewable and Sustainable Energy
Reviews, 56, 246-260.
18. Bailek, N., Bouchouicha, K., Al-Mostafa, Z., El-Shimy, M., Aoun, N., Slimani, A. and Al-Shehri, S.
(2018) A New Empirical Model for Forecasting the Diffuse Solar Radiation over Sahara in the
Algerian Big South. Renewable Energy, 117, 530-537.
14
19. Khalil, S.A. and Shaffie, A.M. (2013) A Comparative Study of Total, Direct, and Diffuse Solar
Irradiance by Using Different Models on Horizontal and Inclined Surfaces for Cairo, Egypt. Renewable
and Sustainable Energy Reviews, 27, 853-863
20. Akpootu, D. O., & Iliyasu, M. I. (2015). A comparative study of some meteorological parameters for
predicting global solar radiation in Kano, Nigeria based on three variable correlations. Advances in
Physics Theories and Applications, 49, 1–9.

15