Predictive modeling and optimization of a waste cooking oil biodiesel/diesel powered CI engine: an RSM approach with central composite design Authors: Medhat Elkelawy, E A El Shenawy, Hagar Alm-Eldin Bastawissi, Mahmoud M Shams, Elumalai P V, Dhinesh Balasubramanian, Vivek Anand, Mamdooh Alwetaishi Published: 2024 ABSTRACT AbstractDue to the restrictions of the diesel engine emissions and the massive demand of energy, the fossil diesel fuel has been consumed quickly and the resources cannot suffice the demand. Alternative fuels that include bio alcohols, hydrogen and biodiesel can make up the diesel fuel depletion. Biodiesel is convenient for diesel engine operation due to its properties like fossil diesel properties. Response surface methodology is a statistical approach for responses prediction and optimization using definite number of experiments to provide time and cost. This study aims to predict and optimize the performance and emission attributes; of diesel engine has single cylinder and operates at 1400 rpm constant speed fuelled with pure diesel fuel or diesel fuel blended with waste cooked oil (WCO) biodiesel at different blending ratio by using response surface methodology (RSM). The influences of the independent variables that are WCO biodiesel blend percentages and the diesel engine load values on the responses that are predicted and optimized. The WCO biodiesel/ diesel fuel blend percentages are pure diesel fuel without biodiesel (B0), 40% WCO biodiesel with 60% diesel fuel (B40) and 80% WCO biodiesel with 20% diesel fuel (B80). The experiments are performed using diesel engine runs at 1400 rpm constant speed, at varying diesel engine loads are zero, 4 and 8 kW. The design of experiments (DOE) is attempted using central composite design (CCD). The RSM model is a nonlinear model developed according to the independent variables and the responses. The homogeneity between the independent variables is studied to predict and optimize their influences on the diesel engine performance and emission attributes. The RSM model is validated according to the coefficients of regression are R2, R2 adjusted and the R2 predicted that prove the satisfaction of the results. From the experiments it is observed that diesel engine performance and emissions attributes are enhanced by increasing the diesel engine load value and increasing the percentage of WCO biodiesel blending ratio compared to pure diesel only like NOx emissions which reduced from 1200 ppm to 900 ppm at the same engine load due to the reduced combustion temperatures using WCO biodiesel. According to the response optimizer tool, the optimal responses are 17.11% for the BTE, 658.9 ppm for the NOx emissions and 1.93% for CO2 emission at independent variables are 2.6667 kW diesel engine load and 100% pure diesel fuel.Keywords: Diesel fuel, WCO biodiesel, RSM, Diesel engine, Emissions, Performance, CCDSubject terms: Chemistry, Engineering Introduction Energy is one of the main resources for raising any country economy. The need for energy is growing up due to facing the population needs. The energy is consumed often in the industry, agricultural and transportation. The internal combustion engines are one of the energy production resources that run using fossil fuels like diesel fuel or the gasoline. Due to using of fossil fuels, the internal combustion engines produce excessive amounts of greenhouse emissions like CO2emissions that estimated billion of tons releasing through the current decade. Also, it is expected that the temperature of the atmosphere will rise 2 °C at 20501. Diesel fuel is one of the non-renewable resources that can be depleted in addition to cause damages to the environment and the humankind healthy2,3. The massive consumption for the non-renewable fuel resources can cause a global economy crisis4,5. Governments have carried out restrictive regulations to organize the pollution resources that represented by the dependence on the alternative fuels6. The fossil fuel can be substituted by the alternative fuels that characterized by having properties similar to fossil fuels in addition high oxygen content, high cetane number, biodegradable, producing high performance attributes for the diesel engines and less emissions7. However, the economical point of view for using the waste cooking oil as a new renewable sours of energy presented a new promising direction for applying the sustainability in the field of combustion control8. Also, the optimal operating conditions of direct injection diesel engine fueled with different biodiesel blends has been achieved with numerical and experiment techniques9,10. In this regard, the diesel engine load variation with respect to biodiesel up to 35% by volume and different exhaust gas recirculation have been studied by RSM11. The obtained results elucidate the limitation and best operating parameters that can be conceded to obtain the maximum BTE and minimize the engine emissions. From the alternative fuels types, biofuels which include biodiesel, biogas, syngas and bio-alcohols4,12. Biodiesel can be produced from many feed stocks that are edible, nonedible, animal fats or other resources13. The edible oil resources that can be used in biodiesel production may restrict the food chain continuity. Sources of waste cooked oil (WCO) for production WCO biodiesel are industrial food processing which represent 32%, various restaurants represent 47% and householders represent 18%14. WCO is one of biodiesel resources but must pass through treatment before using as alternative fuel in diesel engines due to its variant properties compare to fossil diesel fuel15. WCO biodiesel can be specified by oxygenated, has high cetane number, nontoxic, biodegradable and less sulfur content16. There are many methods for converting WCO to WCO biodiesel for diesel engine combustion purposes are transesterification, micro-emulsion, dilution and pyrolysis17. Transesterification is a chemical process that was attempted in that manuscript to convert WCO to WCO biodiesel and glycerol by using methanol and NaOH that used as a catalyst18. Different percentages of WCO biodiesel blends were used to investigate the attributes of partially premixed charge compression ignition (PPCCI) engine. The PPCCI mode was operated after starting the diesel engine at the conventional mode to avoid starting at low temperatures. The results show reductions at the in-cylinder peak pressures and at the heat release rates. Also it was observed enhancements in the BTE percentages and at most of emissions compared with operation at direct injection mode19. Tobacco seed oil biodiesel was used with diesel fuel to investigate the characteristics for Kirloskar diesel engine operating at varying loads, constant speed and injection pressure that were 1500 rpm and 200 bar, respectively. From the results, it was proved that the tobacco seed oil biodiesel blends can enhance the characteristics of the diesel engine20. The Usage of WCO biodiesel blends added by different concentrations of nanoparticles like CeO2 or MgO was studied. All experiments proved that the blends B10CeO2 and B20CeO2 have enhanced the performance and emission characteristics such as BTE, BSFC, NOx, UHC and CO221. Previously, investigation and optimization of the diesel engine attributes consume more time, cost, and vast number of experiments. So, the attributes can be predicted and optimized by using response surface methodology (RSM) approach to reduce time, cost and number of experiments by selecting the most influenced independent variables on the responses22. RSM is a statistical technique used to develop models for predicting and optimizing the responses11. The design of the experiments (DOE) used to identify the number of experiments with the values of the independent variables to get the responses values. In this manuscript the independent variables are the diesel engine load and the WCO biodiesel blend percentages, and the responses are the brake thermal efficiency (BTE) percentage, nitrogen oxides (NOx) emissions and carbon dioxide (CO2) emission. The attributes of reactivity-controlled compression ignition (RCCI) diesel engine mode were investigated. The RSM was applied by Box-Behnken to predict the optimal independent variables and all responses23. TiO2 nanoparticles at 25, 50, and 75 ppm concentrations were mixed with hemp seed oil biodiesel blend at 30% to perform experiments using diesel engine to identify the optimal concentration of the nanoparticles in the blend. The optimal independent variables were 75 ppm concentration of TiO2 and 2 kW engine load; and the responses at these conditions were 27.942% for the BTE, 1081.51 g/kWh for the brake specific fuel consumption, 0.057% for CO emission, 40.293 ppm for the unburned hydrocarbons, 257.3742 ppm for the NOxemissions and 0.7064% for the smoke emissions24. The manuscript objective is predicting and optimizing for the attributes of diesel engine performance and emissions. The diesel engine has single cylinder and operates at constant speed fuelled by pure diesel fuel and high blends percentages of WCO biodiesel blends at variant engine loads. The novelty of the manuscript is studying the effects of high percentages of WCO biodiesel blends and the maximum allowable engine load on the diesel engine attributes. The design of the experiments is attempted according to central composite design technique at various WCO blend percentages and engine load by using MINITAB software application to get responses with accurate values. The independent variables are the diesel engine load value and the WCO biodiesel blend percentages. The responses for the diesel engine performance and emissions are the BTE, NOx emissions and CO2 emission. The experimental setup and procedure The engine setup ZS1115NM diesel engine is used to perform all experiments for predicting and optimizing all diesel engine attributes. The diesel engine that operates at 1400 rpm connected with ATE-160LC hydraulic dynamometer for measuring the loads. Table 1 shows the specifications of the diesel engine and Table 2 shows the hydraulic dynamometer specifications. Fuel injection system is used for injecting all fuel blends. Temperature thermocouples are used for measuring lubrication oil, exhaust gas and cooling water temperatures and data loggers are used to view the temperatures. The experimental test rig and the schematic diagram for the diesel engine, hydraulic dynamometer and all equipments are shown in Figs. 1 and 2, respectively. The diesel engine is supplied with gas board 5020-gas analyzer for measuring all emissions emitted from the combustion CO, CO2, NOx and UHC. The measuring process can attempt by putting the gas analyzer probe into the exhaust outlet and the results can be obtained by the screen of the gas analyzer or can be printed. The gas analyzer specifications are introduced in Table 3. Table 1Specifications of ZS1115NM diesel engine.No.SpecificationsValue123456789101112ModelNo. of CylindersDisplacementBoreStrokeCompression ratioCooling systemMaximum powerMaximum speedNet weightLubrication systemStarting systemZS1115NMSingle cylinder1.195 L115 mm115 mm17.6Evaporation or radiator18 kW2200 rpm185 KgPressure /splashElectrical start Specifications of ZS1115NM diesel engine. 1 2 3 4 5 6 7 8 9 10 11 12 Model No. of Cylinders Displacement Bore Stroke Compression ratio Cooling system Maximum power Maximum speed Net weight Lubrication system Starting system ZS1115NM Single cylinder 1.195 L 115 mm 115 mm 17.6 Evaporation or radiator 18 kW 2200 rpm 185 Kg Pressure /splash Electrical start Table 2Specifications of ATE-160 LC hydraulic dynamometer.No.SpecificationsValue1234567ModelLoad cell capacity.Length of the calibration armConnection with diesel engineMechanism of loadAbsorptionRPM sensingATE-160 LC0–350 Kg0.7645 mUsing half couplingWeight cell and indicator for torqueHydraulic60 tooth gear Specifications of ATE-160 LC hydraulic dynamometer. 1 2 3 4 5 6 7 Model Load cell capacity. Length of the calibration arm Connection with diesel engine Mechanism of load Absorption RPM sensing ATE-160 LC 0–350 Kg 0.7645 m Using half coupling Weight cell and indicator for torque Hydraulic 60 tooth gear The experimental test rig. Schematic diagram for the diesel engine and its accessories. Table 3Specifications of gas board 5020 gas analyzer.No.SpecificationsValue1ModelGas board 50202TechnologyCO2, CO and UHC using non-dispersive infrared technology (NDIR),O2 and NOx using emission control device (ECD)3Resolution(CO2: 0.01%) – (CO: 0.01%) (UHC: 1ppm) – (NOx: 1ppm)4Relative error(CO2: ±4%) - (CO: ±3%)(UHC: ±5%) - (NOx: ±5%)5Absolute error(CO2: ±0.4%) - (CO: ±0.06%) (UHC: ±12 ppm) - (NOx: ±25 ppm)6Warm-up time10 min7Operating temperature(0℃ − 40℃)8Dimension(260 × 180 × 360) mm Specifications of gas board 5020 gas analyzer. CO2, CO and UHC using non-dispersive infrared technology (NDIR), O2 and NOx using emission control device (ECD) (CO2: ±4%) - (CO: ±3%) (UHC: ±5%) - (NOx: ±5%) Preparation of WCO biodiesel and fuel blends Waste cooked oil must undergo to a chemical treatment to use as a biodiesel has specifications near to the specifications of the diesel fuel, due to the soaring prices, and overcome the depletion. Transesterification is the chemical process that can be attempted at the laboratory of fuels as indicated in Fig. 3to produce WCO biodiesel by converting WCO to free fatty acid methyl ester and glycerol that can be used in cosmetics or production of soap according three steps are the reaction, separation and the washing18. The first step is the reaction by mixing the WCO with alcohol that is methanol and catalyst like NaOH then heat the mixture for 60 min at 65 °C. The second step is the separation by leaving the mixture for 24 h to the WCO biodiesel separate from the glycerol25, 26. The third step is washing step by using water at the boiling temperature. The produced WCO biodiesel can use as an alternative fuel with fossil diesel fuel to enhance the attributes of the diesel engine. The specifications of the WCO biodiesel and fossil diesel fuel are introduced in Table 4. WCO biodiesel production method by transesterification. Table 4Specifications of pure diesel fuel and WCO biodiesel.Fuel itemsUnitsDiesel fuelWCO BiodieselCetane number----5568Density(Kg/m³)831874Heating rate(MJ/kg)42.1239.52Viscosity(mm²/s)2.383.54Flashpoint(°C)46159Oxygen content(Wt.%)09.414Carbon residue(Wt.%)04% Specifications of pure diesel fuel and WCO biodiesel. The uncertainty test Due to the experiments have the tendency to make errors; the uncertainty test is especially important to determine the truth of the experiment results. The uncertainty test related to some factors like the instruments, surrounding environment, the techniques of measurements and other factors are calculated according to square root method are indicated in Table 5. Table 5The uncertainty values for the following factors.EquipmentUncertaintyNOx emissions± 0.5%CO2 emission± 0.5%Temperature transmitter± 1 deg.Brake thermal efficiency± 1.66% The uncertainty values for the following factors. Design of experiments RSM is a technique utilized in prediction the diesel engine attributes in addition and optimization. Firstly, it is important to get relations between the independent variables and the responses. Design of the experiments (DOE) applied to select various values of the independent variables to calculate the responses. If the relations are linear, the DOE can be applied by full factorial design (FFD). For non-linear or curvature relations as in Eq. (1) the DOE can be applied by central composite design (CCD) or Box Behnken design using statistical software like Minitab (Minitab Inc.). (i) is the linear coefficient, (k) is the independent variable numbers, (j) is the quadratic variable, (y) is the response, (β) is the regression coefficient and (ε) is the experimental error. In the current manuscript, the independent variables are the engine load in (kW) and the blend percentage of WCO biodiesel in (%). The independent variables with their levels are indicated in Table 6. After selecting the independent variables, their levels, and the CCD technique to design the experiments now the thirteen experiments can be performed and get the responses as mentioned in Table 7. The relationships between the independent variables and the responses are represented by three quadratic regression equations are Eqs. (2, 3 and 4) which are derived from the experimental data. Table 6The independent variables with their levels.The independent variableLowMediumHigh Engine load (A) (kW) 048 WCO Blend percentage (B) (%) 04080 The independent variables with their levels. Table 7The data of the experiments for the responses and the independent variables.Run OrderEngine loadBlend %ExperimentalPredictedBTENOxCO2BTENOxCO2 1 2 3 4 5 6 7 8 9 10 11 12 13 44444080804484080040400080804040404023.3420.7022.5023.3022.700.0030.930.0027.900.0023.1022.8028.368056208207957553151402250100428477076312502.712.602.202.652.411.373.651.514.161.432.502.463.8022.775521.476223.086222.775522.7755-0.117330.4611-0.212327.33610.329522.775522.775529.3929773.00621.00842.00773.00773.00296.421398.58241.921011.08310.67773.00773.001246.332.517932.645172.295172.517932.517931.338253.586581.503254.121581.468512.517932.517933.90184 The data of the experiments for the responses and the independent variables. 1 2 3 4 5 6 7 8 9 10 11 12 13 4 4 4 4 4 0 8 0 8 0 4 4 8 40 80 0 40 40 0 0 80 80 40 40 40 40 23.34 20.70 22.50 23.30 22.70 0.00 30.93 0.00 27.90 0.00 23.10 22.80 28.36 805 620 820 795 755 315 1402 250 1004 284 770 763 1250 2.71 2.60 2.20 2.65 2.41 1.37 3.65 1.51 4.16 1.43 2.50 2.46 3.80 22.7755 21.4762 23.0862 22.7755 22.7755 -0.1173 30.4611 -0.2123 27.3361 0.3295 22.7755 22.7755 29.3929 773.00 621.00 842.00 773.00 773.00 296.42 1398.58 241.92 1011.08 310.67 773.00 773.00 1246.33 2.51793 2.64517 2.29517 2.51793 2.51793 1.33825 3.58658 1.50325 4.12158 1.46851 2.51793 2.51793 3.90184 Where (A) is the diesel engine load and (B) is WCO biodiesel blend percentage. The analysis of variance (ANOVA) is used to indicate the quadratic regression models significance and the interaction between the independent variables on the responses. The degree of freedom (DF) for any model is the number of experiments subtracted by one. The adjusted sum of squares is represented by (Adj SS), and the adjusted mean of squares is represented by (Adj MS) and calculated by dividing the sum of squares for the model to the degree of freedom for the model. The P-value and the F-value are used to indicate the regression model significance where the F-value is the Fisher test value that can describe the variance measurements according to the mean basing on the ratio between the mean square of variance to the error. The independent variables significance can be determined if the P-value lowering 0.05. The analysis of variance for the BTE, NOx and CO2 are indicated in Tables 8, 9 and 10 respectively the results of F-value are 607.66, 548.43 and 129.57 and the P-value are 0.000 for the BTE, NOx and CO2 respectively prove that the statistically significance of the regression models for the responses. The coefficients of determination that are R2, R2 adjusted and R2 predicted with high percentages that describe the good fitting between the experimental and the predicted results. Table 8The ANOVA for the BTE.BTE (%)Degree of freedomAdj MSAdj SSFisher test valueP-Value Model Linear A B Square A×A B×B 2-way interaction A×B Error Lack of fit Pure error Total 5 2 1 1 2 1 1 1 1 7 3 4 12 297.20 635.45 1267.02 3.89 106.39 173.00 0.67 2.30 2.30 0.49 1.03 0.08 1485.98 1270.90 1267.02 3.89 212.78 173.00 0.67 2.30 2.30 3.42 3.09 0.33 1489.41 607.66 1299.26 2590.57 7.95 217.53 353.71 1.38 4.69 4.69 12.33 0.000 0.000 0.000 0.026 0.000 0.000 0.279 0.067 0.067 0.017 R2 = 99.77%% R2(adjusted) = 99.61% R2(predicted) = 98.46% The ANOVA for the BTE. Model Linear A B Square A×A B×B 2-way interaction A×B Error Lack of fit Pure error Total 5 2 1 1 2 1 1 1 1 7 3 4 12 297.20 635.45 1267.02 3.89 106.39 173.00 0.67 2.30 2.30 0.49 1.03 0.08 1485.98 1270.90 1267.02 3.89 212.78 173.00 0.67 2.30 2.30 3.42 3.09 0.33 1489.41 607.66 1299.26 2590.57 7.95 217.53 353.71 1.38 4.69 4.69 12.33 0.000 0.000 0.000 0.026 0.000 0.000 0.279 0.067 0.067 0.017 R2 = 99.77%% R2(adjusted) = 99.61% R2(predicted) = 98.46% Table 9The ANOVA for the NOx.NOx (ppm)Degree of freedomAdj MSAdj SSFisher test valueP-Value Model Linear A B Square A×A B×B 2-way interaction A×B Error Lack of fit Pure error Total 5 2 1 1 2 1 1 1 1 7 3 4 12 283,858 693,235 1,313,208 73,262 2550 84 4757 27,722 27,722 518 596 459 1,419,292 1,386,470 1,313,208 73,262 5100 84 4757 27,722 2722 3623 1788 1835 1,422,915 548.43 1339.37 2537.19 141.55 4.93 0.16 9.19 53.56 53.56 1.30 0.000 0.000 0.000 0.000 0.046 0.700 0.019 0.000 0.000 0.390 R2 = 99.75%% R2(adjusted) = 99.56% R2(predicted) = 98.71% The ANOVA for the NOx. Model Linear A B Square A×A B×B 2-way interaction A×B Error Lack of fit Pure error Total 5 2 1 1 2 1 1 1 1 7 3 4 12 283,858 693,235 1,313,208 73,262 2550 84 4757 27,722 27,722 518 596 459 1,419,292 1,386,470 1,313,208 73,262 5100 84 4757 27,722 2722 3623 1788 1835 1,422,915 548.43 1339.37 2537.19 141.55 4.93 0.16 9.19 53.56 53.56 1.30 0.000 0.000 0.000 0.000 0.046 0.700 0.019 0.000 0.000 0.390 R2 = 99.75%% R2(adjusted) = 99.56% R2(predicted) = 98.71% Table 10The ANOVA for the CO2.CO2 (%)Degree of freedomAdj MSAdj SSFisher test valueP-Value Model Linear A B Square A×A B×B 2-way interaction A×B Error Lack of fit Pure error Total 5 2 1 1 2 1 1 1 1 7 3 4 12 1.83554 4.53271 8.88167 0.18375 0.03904 0.07725 0.00630 0.03422 0.03422 0.01417 0.01115 0.01643 9.17771 9.06542 8.88167 0.18375 0.07807 0.07725 0.00630 0.03422 0.03422 0.09916 003344 0.06572 9.27688 129.57 319.97 626.96 12.97 2.76 5.45 0.44 2.42 2.42 0.68 0.000 0.000 0.000 0.009 0.131 0.052 0.526 0.164 0.164 0.609 R2 = 98.93% R2(adjusted) = 98.17% R2(predicted) = 96.33% The ANOVA for the CO2. Model Linear A B Square A×A B×B 2-way interaction A×B Error Lack of fit Pure error Total 5 2 1 1 2 1 1 1 1 7 3 4 12 1.83554 4.53271 8.88167 0.18375 0.03904 0.07725 0.00630 0.03422 0.03422 0.01417 0.01115 0.01643 9.17771 9.06542 8.88167 0.18375 0.07807 0.07725 0.00630 0.03422 0.03422 0.09916 003344 0.06572 9.27688 129.57 319.97 626.96 12.97 2.76 5.45 0.44 2.42 2.42 0.68 0.000 0.000 0.000 0.009 0.131 0.052 0.526 0.164 0.164 0.609 R2 = 98.93% R2(adjusted) = 98.17% R2(predicted) = 96.33% The difference between the actual results and the predicted results called residuals. Figure 4 shows the residuals normal probabilty plots for the BTE, NOx and CO2 respectively which the residuals are represented by straight line that prove that the residuals for all engine attributes follow normal distributions. Also, the relation between the actual results and the predicted results are represented in Fig. 5 for the BTE, NOx and CO2. the values of R2 and the values of R2 adjusted nearly equal to one are evidence for the models to get accurate estimation for the responses. The normal probability plots for the residuals of the responses. The relations between the actual and the predicted responses. The experimental setup and procedure The present manuscript can predict and optimize the influences of using the following independent variables are the WCO blend percentages with three levels are 0, 40 and 80% in addition different diesel engine loads with three levels also are 0, 4 and 8 kW on the diesel engine attributes of diesel engine operates at 1400 rpm constant speed. The responses are the BTE, the NOx and the CO2 emission. RSM is a statistical technique used in prediction and optimization the attributes of the diesel engine by reduction the numbers of the experiments, time and cost using RSM software application (MINITAB17). The influences of the independent variables with their levels diesel engine load values and the WCO blend percentages on the brake thermal efficiency (BTE) are indicated in Fig. 6. As seen, there is sharp increasing in the BTE just with increasing in the values of the load due the more conversion of fuel energy to mechanical work and so increasing in the brake power and subsequently increasing in the BTE. In addition, slow increasing in the BTE just with increasing in the WCO biodiesel blend percentages from 0% that is pure diesel fuel to 80% WCO biodiesel blend. The causes are the raising in the oxygen content by using the WCO biodiesel in addition the high cetane number by increasing the WCO blend percentage. indicates the surface and the contour plots for the influences of engine load and WCO biodiesel blend percentages on the BTE. The influences of the diesel engine load values and the WCO biodiesel blend percentages on the NOx emissions are indicated in Fig. 7. As seen, there are increasing in the NOx emissions by increasing the values of the engine load due to the complete combustion and the high combustion temperatures that is the main cause of increasing NOx emissions. In addition, the NOx emissions decrease gradually with converting the fuel from pure diesel to WCO biodiesel blend percentages due to the low combustion temperatures that produced with increasing the percentage of the WCO biodiesel percentages. indicates the surface and contour plots for the influences of engine load values and WCO biodiesel blend percentages on the NOx values. The values of the load and the percentages of WCO biodiesel blend influences on the CO2 emission percentages are indicated in Fig. 8. As seen, there is increasing in the CO2 emission by increasing in the values of the load because the complete combustion which produce high combustion temperatures and subsequently converting all CO emission to CO2 emission. In addition, the percentages of CO2 emissions are increased with converting the fuel from pure diesel to gradual percentages of WCO biodiesel blend because the WCO biodiesel high oxygen level that complete the combustion and convert CO emission to CO2 emission. indicates the surface and contour plots for the influences of the load in addition the WCO biodiesel blend percentages on the CO2 emission. The responses optimization and desirability The optimizer tool at MINITAB 17 software application is applied to obtain the optimum values for the independent variables to optimize the output responses. To get the optimum values a group of thirteen experiments was performed and Fig. 9 shows the desirability for each independent variable, output response with its individual desirability and the composite desirability (D). As indicated the best values for the independent variables are 2.66 kW for the engine load and 0% for the WCO biodiesel blend percentage. The optimum values for the output responses are 17.11% for the BTE, 658.9 ppm for the NOx emissions and 1.93% for the CO2 emission. The composite desirability has 0.65-rounded value and the individual desirability values for the BTE, NOx emissions and CO2 emission are 0.55, 0.64 and 0.79, respectively. The optimization plot for optimizing the engine attributes. Conclusions and future works The aim of this manuscript is predicting and optimizing the diesel engine attributes fuelled with blends of WCO biodiesel according to response surface methodology. RSM can save time, cost, and reduction number of performing experiments. The responses are the BTE, NOx emissions and CO2 emission, for the independent variables are the engine load value and WCO biodiesel blend percentages. The design of the experiments according to central composite design through two factors each has three levels, and the optimization is achieved using optimizer tool by MINITAB 17 software application and the following points include the conclusions. Enhancements in the BTE, NOx emissions and CO2 emissions by using different values of the diesel engine loads and different percentages of WCO biodiesel blend percentages due to the improved fuel properties like high oxygen content, high cetane number the combustion completion. Enhancements in the BTE, NOx emissions and CO2 emissions by using different values of the diesel engine loads and different percentages of WCO biodiesel blend percentages due to the improved fuel properties like high oxygen content, high cetane number the combustion completion. The values for the regression model coefficients [R2, R2 (adjusted) and R2 (predicted)] are more than 95% that evidence for the models are fitted with the experimental data. The values for the regression model coefficients [R2, R2 (adjusted) and R2 (predicted)] are more than 95% that evidence for the models are fitted with the experimental data. The optimizer tool using MINITAB 17 software is used to maximize the BTE, minimize NOx emissions and CO2 emission. The optimal for the responses are 17.11% for the BTE, 658.9 ppm for the NOx emissions and 1.93 for the CO2 emission, while the values for the independent variables are 2.66 kW for the engine load and 100% diesel fuel in addition the composite desirability (D) has value 0.65. The optimizer tool using MINITAB 17 software is used to maximize the BTE, minimize NOx emissions and CO2 emission. The optimal for the responses are 17.11% for the BTE, 658.9 ppm for the NOx emissions and 1.93 for the CO2 emission, while the values for the independent variables are 2.66 kW for the engine load and 100% diesel fuel in addition the composite desirability (D) has value 0.65. The recommendations for the future works are: Using several types of nanoparticles that can be added to the fuel blend to enhance the physical and chemical properties for the fuel and enhance the performance and emissions attributes for the diesel engine. Using several types of nanoparticles that can be added to the fuel blend to enhance the physical and chemical properties for the fuel and enhance the performance and emissions attributes for the diesel engine. Using diverse types of biodiesels and study their influences on the diesel engine attributes. Using diverse types of biodiesels and study their influences on the diesel engine attributes. Nomenclature BTE brake thermal efficiency BSFC brake specific fuel consumption CCD central composite design CO2 carbon dioxide NOx oxides of nitrogen PPM parts per millions PPCCI partially premixed charge compression ignition RSM response surface methodology UHC unburned hydrocarbon WCO waste cooked oil Acknowledgements The authors extend their appreciation to Taif University, Saudi Arabia, for supporting this work through project number (TU-DSPP-2024-32). The authors thankfully recognize the financial of the Tanta University Research Fund. This experimental work was sponsored by the Tanta University Research Fund under the research grant (code: tu: 02-19-01). The authors would like to acknowledge the management of the Mepco Schlenk Engineering College, Sivakasi, Tamil Nadu, India for encouraging and supporting the research work. Authors’ contributions M.E, E.A.E.S, H.A.E.B wrote the main manuscript text and M.M.S, E.P.V, D.B. V.A, M.A, prepared the experimental work. All authors reviewed the manuscript. Data availability The datasets used and/or analysed during the current study available from the corresponding author on reasonable request. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Medhat Elkelawy, Email: medhatelkelawy@f-eng.tanta.edu.eg. Elumalai P. V., Email: elumalai@adityauniversity.in. Dhinesh Balasubramanian, Email: Dhineshbala91@mepcoeng.ac.in. References Associated Data Data Availability Statement The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.