Gasoline Samples Research Articles

Smoke point prediction of oxygenated fuels using neural networks

Smoke point (SP) is an important fuel property that characterizes the propensity of aviation jet fuels and kerosene to form soot. In the present study, an artificial neural network (ANN) model based on the artificial intelligence principle was developed to predict the SP of fuels with oxygenates (ethers and alcohols) and hydrocarbons (e.g., paraffins, olefins, naphthenes, aromatics, and their blends). An experimental dataset of 366 fuel mixtures comprising 113 pure compounds, 8 types of gasoline and diesels, and 245 fuel blends was used to improve the SP-prediction performance of the ANN model. One hundred and ten of the fuel blends used were various jet fuel surrogates collected from the literature. Experimental SP of 40 new mixtures consisting of diethyl ether, dibutyl ether, dioctyl ether, diphenyl ether, and methyl tert-butyl ether along with gasoline and diesel was measured as part of this work. The molecular composition of the fuels was expressed as the weight % of eight constituent functional groups present in the fuel. These functional groups along with two additional parameters (branching index and molecular weight) were provided as ten inputs to the model. The functional groups present in the gasoline and diesel samples were determined using high-resolution 1H nuclear magnetic resonance spectroscopy. The data were randomly split into three sets: training (70 %), validation (15 %), and test (15 %). The model was initially trained and validated simultaneously, and then; it was tested. A positive linear relation was observed between the measured and predicted SPs, as indicated by a correlation coefficient of 0.98. The mean absolute error of the predicted SP was 4.5. Results showed that the SP of the fuels depended on the aforementioned parameters that served as inputs for the model. The proposed model can be used to predict the SP of pure and blended form fuels containing the aforementioned functional groups.

Quality classification of gasoline samples based on their aliphatic to aromatic ratio and analysis of PONA content using genetic algorithm based multivariate techniques and ATR-FTIR spectroscopy

In this study, attenuated total reflectance fourier transform mid-infrared (ATR-FTIR) spectral information was used to classify gasoline samples according to their aliphatic to aromatic ratio using data driven soft independent modeling of class analogy (dd-SIMCA). The ATR-FTIR spectral data were also used to determine the paraffin, olefin, naphthene and aromatic (PONA) contents in gasoline samples using support vector machine regression (SVMR) model. For comparison, partial least squares discriminant analysis (PLS-DA) and partial least square regression (PLSR) were also performed for classification and prediction purposes, respectively. Genetic algorithm (GA) variable selection method was applied to achieve a better separation of the classes and select characteristic variables closely related to composition differences. The results indicated that GA-dd-SIMCA and GA-SVMR can provide better classification and prediction performances by comparison with GA-PLS-DA and GA-PLSR methods, respectively. For supervised dd-SIMCA discrimination method with GA variable selection, 100% correct classification for three different classes of the gasoline was obtained, each of which has different aliphatic/aromatic ratio. In addition, the optimal GA-SVMR model showed the determination coefficient (R2) of 0.985, 0.986, 0.960 and 0.981 and root mean square error of prediction (RMSEP) were 2.050, 1.243, 2.189 and 3.324 for parrafins, olefins, naphthenes and aromatics, correspondingly. FTIR spectroscopy coupled with multivariate data analysis procedures provides rapid, efficient, and cost-effective quality assurance methods to characterize physicochemical properties of gasoline samples.

SBA-15 supported benzoxazolium-based ionic liquids: Synthesis, characterization, and application in the adsorptive desulfurization

In this study, organosilane-modified S-aryl/alkyl-substituted benzoxazolium-based ionic liquids ([S-RBOx][Cl]; R = benzyl or methyl) were synthesized and immobilized on SBA-15 silica support to afford supported ionic liquids (SILs). The structure elucidation of SILs was done by using SEM, SAXS, BET, TEM, EDS, and NMR spectroscopy. Moreover, the adsorbent role of SILs and SBA-15 was studied in the adsorptive desulfurization by taking model and real gasoline samples. In contrast to SBA-15, both SIL adsorbents were found more efficient because of the π-π interactions, induced between supported [S-RBOx][Cl] and thiophene compounds. The Freundlich adsorption isotherm was found more suitable to fit the experimental adsorption data in comparison to the Langmuir isotherm. SBA-15@[S-BzBOx][Cl] was obtained with maximum adsorption capacity (328 mg/g) in comparison to SBA-15 (109.1 mg/g) and SBA-15@[S-MeBOx][Cl] (256.6 mg/g). The maximum adsorption was obtained by SBA-15@[S-BzBOx][Cl] at 45 °C within 100 min by using an adsorbent quantity of 600 mg from model gasoline having a sulfur content of 250 ppm and provided removal of TS, TN, and DBT with the percentage of 57.2%, 63.6%, and 78.9% respectively. All the SILs were reused ten times and obtained with almost constant adsorption efficiency.

Profiling Olefins in Gasoline by Bromination Using GC×GC-TOFMS Followed by Discovery-Based Comparative Analysis.

An analytical workflow for the analysis of olefins in gasoline that combines selective bromination and comprehensive two-dimensional (2D) gas chromatography time-of-flight mass spectrometry (GC×GC-TOFMS) with discovery-based analysis is reported. First, a standard mix containing n-alkanes, 1-alkenes, and aromatic species was brominated and quantified using % reacted as a metric for each compound class, defined as the difference in the total peak area between the brominated and original samples normalized to the original sample. The average % reacted (1 s.d.) values were -1.45% (2.8%) for the alkanes, 99.5% (0.4%) for the alkenes, and 6.7% (11.6%) for the aromatics, demonstrating excellent selectivity toward the alkenes with only minor aromatic bromination. The bromination chemistry was then applied to gasoline, followed by GC×GC-TOFMS analysis of the original and brominated gasoline. This GC×GC-TOFMS data set was then submitted to the supervised discovery tool tile-based F-ratio analysis (FRA), which reduced the large data set to only the chromatographic regions which distinguish between the original and brominated gasoline samples. FRA discovered 314 hits, 56 of which were determined statistically significant using combinatorial null distribution analysis (CNDA), a permutation-based significance test. Since the brominated olefins elute in an uncrowded region of the 2D chromatogram and have no signal in the original sample, their discoverability was greatly increased relative to the original olefins. By combining the information gained from brominated olefin standards and the structured patterns of the GC×GC separations, the top hits were identified as the dibromoalkane products of mono-olefins, with five C5 mono-olefins identified on a species level. The analytical workflow has broad implications for using selective reaction chemistries to facilitate supervised discovery by targeting desired compound classes.

A Predictive Approach towards Using PC-SAFT for Modeling the Properties of Shale Oil.

Equations of state are powerful tools for modeling thermophysical properties; however, so far, these have not been developed for shale oil due to a lack of experimental data. Recently, new experimental data were published on the properties of Kukersite shale oil, and here we present a method for modeling the properties of the gasoline fraction of shale oil using the PC-SAFT equation of state. First, using measured property data, correlations were developed to estimate the composition of narrow-boiling-range Kukersite shale gasoline samples based on the boiling point and density. These correlations, along with several PC-SAFT equations of the states of various classes of compounds, were used to predict the PC-SAFT parameters of aromatic compounds present in unconventional oil-containing oxygen compounds with average boiling points up to 180 °C. Developed PC-SAFT equations of state were applied to calculate the temperature-dependent properties (vapor pressure and density) of shale gasoline. The root mean square percentage error of the residuals was 13.2%. The average absolute relative deviation percentages for all vapor pressure and density data were 16.9 and 1.6%, respectively. The utility of this model was shown by predicting the vapor pressure of various portions of the shale gasoline. The validity of this model could be assessed for oil fractions from different deposits. However, the procedure used here to model shale oil gasoline could also be used as an example to derive and develop similar models for oil samples with different origins.

PRODUCTION AND STUDY OF BIO GAZOLINES

Combustible ethanol was obtained from agricultural waste. Test samplesE0 (petrolgazoline) and E5, E10 and E20 (epanpl/gazoline blend) were prepared and their physical, chemical and technical characteristics were studied. The effect of bioethanol on the ecological compatibility of the automobile engine has been studied. The objects of research were Regular, Premium and Super of Petrol gasoline Samples, as well as a of 10%, 20% and 30%. bioethanol/gazoline blends. The study of the ecological properties of the test additive was carried out at the stand of automobile engines SAK-670 (Germany), on which the engine of a BMW-316 car with coupling sleeve and a transmission assembly is installed. The stand is equipped with brakes and torque sensors, as well as with crankshaft speed and fuel consumption measuring instruments. The test mixtures during an idle running of engine cause a decrease in the CO content in the exhaust gases by 10-16%, and this taking into account their low cost, indicate the prospect of expanding the use of fuel ethanol.

Development of lower flammability limits testing apparatus for gasoline vapor – Air mixture using an internal vaporizer at ambient conditions

Fire accidents at public fuel stations in Indonesia have urged the need to characterize the causes of these fires. In the gas station area, the formation of gasoline vapor is the main concern of this research. Therefore, this study aims to analyze lower flammability limits (LFLs) of gasoline vapor-air mixtures of the selected gasoline samples. The LFL of gasoline vapor was characterized using a 1.2 L vertical tube integrated with an internal evaporator and a high voltage electric lighter with the energy of 10 J. The tube is made of glass with a diameter of 80 mm and a height of 300 mm with an open top end. This apparatus can run two test methods of upward and downward propagation. The experiment was conducted at the ambient condition with an evaporation temperature of the sample being 150 °C for the internal evaporator. Four gasoline samples were used with varying octane numbers (RON) obtained from arbitrary public gas stations. For validation, Iso-octane (IO) reference fluid with a purity level of 99.5% was tested for the assessment of LFL. The results with the upward propagation method showed that the LFL volume-based concentration were 1.62%, 1.61%, 1.64%, 1.60% and 1.05%, and downward propagation method were 2.48%, 2.39%, 2.30%, 2.11% and 1.57% for RON_88, RON_90, RON_92, RON_95 and IO_100 samples respectively. This study shows that a downward propagation flame has a higher gas temperature with lower velocity compared to an upward flame, which suggests that lower elevation fuel vapor sources at gas stations have a higher fire potential. This research topic is in an effort to provide input on the fire hazards at gas stations that are specific for tropical conditions with traffic practices such as in Indonesia and other countries that face similar hazards and threats. This study can provide input for further transportation safety policies and research.

Evaluation of the Additive Power of Ethanol Obtained from Angola Grass in Direct Distillation Gasoline Samples, Case Study: Straight Run (SR) Gasoline Produced at Luanda Refinery

Alongside the process of blending naphtha, fuel anhydrous ethyl alcohol is an additive recurrently used to adjust commercial properties such as the octane index of gasolines during its formulation. This additive is environmentally recommended due to the promotion of decarbonization in gasolines that is evidenced in its consumption with the reduction of CO2 emissions into the atmosphere, compared to the consumption of other gasolines. There are several ways to obtain the respective additive, the research leads to obtaining it from the Angolan grass (Brachiaria purpurascens "Forsk" Stapf). Angola grass has a low fiber and protein content, thus becoming with low preference in the choice for animal feed and has a significant content of nonnitrogen extractives giving it a high energy potential to be raw material for biofuel production Through an experimental study and on a laboratory scale, alcohol was produced, then the quantities of the two components of the mixture were determined using additive calculations, and then mixed in samples of Straight run (SR) gasolines direct distillation produced in Luanda Refinery. Next, octane index tests were performed in the final mixture by the RON method following ASTM D22699 and adulteration potencies were verified in the samples. The octane index defined to be reached was 95 octane and it was proven that alcohol produced from Angola grass has the potential to promote improvements in octane index, as we found increases in the octane index of additive gasoline shows. We recorded an improvement of the rate of 91% for sample 1 namely the mixture of anhydrous alcohol of grass and light gasoline (SR) of the Luanda Refinery and finally an improvement in the order of 81% for sample 2 namely the mixture of anhydrous alcohol of heavy gasoline grass of the Luanda Refinery. However, it is demonstrating in the first instance starting methodologies for the additive of gasoline samples, and in the second instance the "effect" of a fuel anhydrous alcohol derived from a "differentiated matter" called Angola grass, when used as an additive in direct distillation gasoline, which is concluded to be positive.

The application of wavelet transform of Raman spectra to facilitate transfer learning for gasoline detection and classification

This study proposed an artificial intelligence (AI) approach for rapid, automatic detection of gasoline samples from different ignitable liquid samples based on their Raman spectra. The system integrated Raman spectroscopy, Raman signal transform, and transfer learning with a convolutional neural network (CNN). A hand-held Raman spectrometer was utilized to collect the learning dataset, including 180 gasoline spectra and 170 non-gasoline spectra from 17 various types of liquid samples. Continuous wavelet transform (CWT) was adopted to transform Raman spectra into image representations to facilitate transfer learning using a pre-trained CNN for image recognition. This approach streamlined AI model training, performance verification, and assessment. Experimental results showed that the CNN model could achieve 100% classification performance in precision, sensitivity, specificity, F1 score, and accuracy. The established AI model could be adopted for the Raman spectrum collected by different accessories, such as using the point-and-shoot attachment of liquid samples in glass vials. The experimental results suggested that CWT processing of Raman scattering signals enabled effective CNN transfer learning. The study demonstrated that CWT could be applied for Raman spectra processing to develop edge AI in forensic field testing of physical evidence.

Paper-Based Optical Nose Made with Bimetallic Nanoparticles for Monitoring Ignitable Liquids in Gasoline.

A paper-based optical nose was fabricated by dropping bimetallic silver and gold nanoparticles on a paper substrate. The nanoparticles were synthesized by both natural (lemon, pomegranate, and orange juices) and chemical (citrate, gallic acid, and ascorbic acid) reducing agents. The performance of the assay was evaluated for identifying gasoline and five ignitable liquids such as diesel, ethanol, methanol, kerosene, and thinner. The interaction of the sensor with sample vapors caused aggregation, consequently changing the color of nanoparticles. The color changes, which were captured by a scanner, represented a specified colorimetric map for each analyte, allowing one to identify the studied fuels. The visual results were confirmed using multivariate statistical analysis such as principal component analysis and hierarchical clustering analysis. Also, partial least-squares regression was used to assist the proposed assay for estimating the amount of studied ignitable liquids as counterfeit species in the gasoline sample. The root-mean-square errors for prediction were 3.4, 2.1, 1.9, 2.0, and 1.7% for diesel, thinner, kerosene, ethanol, and methanol, respectively. Finally, the fabricated sensor indicated high efficiency for the on-site detection of pure industrial gasoline samples from adulterated ones.

Deep eutectic solvent-based air-assisted liquid–liquid microextraction of lead in gasoline samples followed by graphite furnace atomic absorption spectrometry

Deep eutectic solvent-based air-assisted liquid–liquid microextraction of lead in gasoline samples followed by graphite furnace atomic absorption spectrometry

Experimental determination of the lower explosion limit for two gasoline samples

The explosive atmosphere may be caused by flammable gases / vapours or combustible dust. If the amount of the substance, mixed with air, is sufficient, then a source of ignition is needed to cause an explosion. Liquids (for example petrol and other fuels) and solvents from industrial products emit flammable vapours which, when mixed with air, can ignite or explode. At normal temperatures, flammable liquids can emit enough vapours to form combustible mixtures with air, heat, and often thick, black, and toxic clouds of smoke. The behaviour of a fuel-oxidant mixture is characterized by certain explosions parameters, including explosion limits, which characterize the range of concentrations in which combustion propagates at very high speeds. For this work were performed experimental determination of the lower explosion limit for two commercial gasoline samples.

Determination of Oil Pollutants by Microchannel Laser Induced Fluorescence Technology

Laser-induced fluorescence technology is an effective method for detecting oil pollutants. The laser induced fluorescence detection system was developed, and four parallel microchannel sample pools were designed to improve the detection accuracy of fluorescence signals. The grating scanning system is improved by introducing grating feedback link to improve the wavelength scanning accuracy of the system.The experimental results show that the wavelength scanning resolution of the microchannel laser-induced fluorescence detection system is improved by 0.4 nm. Aiming at the defects of the traditional parallel factor algorithm which is sensitive to the sample components, slow operation speed and easy to fall into local convergence, the generalized inverse singular value decomposition is used to improve the parallel factor algorithm. Configure different concentrations of diesel, gasoline and kerosene solution samples. The experimental verification shows that the excitation-emission spectra of each component obtained by the improved parallel factor algorithm are highly consistent with the measured spectral characteristics, indicating that the microchannel laser-induced fluorescence detection system combined with the improved parallel factor analysis method can accurately identify oil pollutants.

Flow visualisation in real-size optical injectors of conventional, additised, and renewable gasoline blends

Research on renewable and alternative fuels is crucial for improving the energy and environmental efficiency of modern gasoline internal combustion engines. To highlight the influence of fuel rheological and thermodynamic properties on phase change and atomisation processes, three types of gasoline blends were tested. More specifically, the campaign comprised a reference gasoline, an ethanol/gasoline blend (10% v/v) representative of renewable fuels, and an additised gasoline sample treated with viscoelasticity-inducing agents. High-speed imaging of the transient two-phase flow field arising in the internal geometry and the near-nozzle spray region of gasoline injectors was performed employing Diffuse Backlight Illumination. The metallic body of a commercial injector was modified to fit transparent tips realising two nozzle layouts, namely a two-hole real size model resembling the Engine Combustion Network spray G injector and an enraged replica with an offset hole. Experiments were conducted at realistic operating conditions comprising an injection pressure of 100 bar and ambient pressures in the range of 0.1–6.0 bar to cover the entire range of chamber pressures prevailing in Gasoline Direct Injection engines. The action of viscoelastic additives was verified to have a suppressive effect on in-nozzle cavitation (6% reduction in cavitation extent) , while also enhancing spray atomisation at flash-boing conditions, in a manner resembling the more volatile gasoline/ethanol blends. Finally, persisting liquid ligaments were found to form after the end of injection for the additised sample, owing to the surfactant nature of the additives.

Comprehensive volatile organic compound measurements and their implications for ground-level ozone formation in the two main urban areas of Vietnam

Volatile organic compounds (VOCs) and oxygenated VOCs (OVOCs) were measured in Ho Chi Minh City (HCMC) and Hanoi, the two largest and most populous cities in Vietnam. The purpose of this study is to better understand the VOC atmospheric composition and their role in ground-level ozone formation. Online measurements of a wide range of VOCs and other pollutants were conducted using numerous instruments during different seasons (dry and rainy) in HCMC and Hanoi (spring). Our results show that the mean mixing ratio of total measured VOCs in Hanoi was 80.8 ± 40.7 ppb (mean ± standard deviation), and was similar to that observed during the rainy season (75.2 ± 44.8 ppb) in HCMC. During the dry season campaign in HCMC, which was coincident with the Hanoi campaign, total VOC was around 50% lower (40.7 ± 19.5 ppb), largely a result of increased planetary boundary layer (PBL) height and the direction of the prevailing wind. VOC profiles in both cities were dominated by alkanes (31–35%) and OVOCs (27–33%) and the proportion of alkenes (13–17%) and aromatics (12–19%) were comparable. Similarities in diurnal variation for most VOC species (except for isoprene) are seen in both cities with two clear peaks during the morning (7:00–8:00 a.m.) and evening (18:00 p.m.) rush hours, as observed for vehicular-combustion tracers (acetylene and CO). Comparisons of the ambient ratios of paired VOCs, namely i/n-pentane, and toluene/benzene, with those reported in motorcycle exhaust, roadside and gasoline samples indicate that motorcycle-related emission is likely a major contributor to VOC pollution. According to the propylene-equivalent concentration (PE conc.) and maximum incremental reactivity (MIR) methods, alkenes and aromatics were determined to be the main contributors to reactivity and ozone potential formation. Furthermore, the initial mixing ratio of VOC species was estimated based on the photochemical age method. The consumed VOCs (initial VOCs minus measured VOCs) has a similar variation trend to ground-level ozone, and a good correlation is observed in HCMC. In contrast, this result was not seen in Hanoi despite relatively high levels of PE conc. and MIR.

Identification and Quantification of Hydrocarbon Functional Groups in Gasoline Using 1H-NMR Spectroscopy for Property Prediction.

Gasoline is one of the most important distillate fuels obtained from crude refining; it is mainly used as an automotive fuel to propel spark-ignited (SI) engines. It is a complex hydrocarbon fuel that is known to possess several hundred individual molecules of varying sizes and chemical classes. These large numbers of individual molecules can be assembled into a finite set of molecular moieties or functional groups that can independently represent the chemical composition. Identification and quantification of groups enables the prediction of many fuel properties that otherwise may be difficult and expensive to measure experimentally. In the present work, high resolution 1H nuclear magnetic resonance (NMR) spectroscopy, an advanced structure elucidation technique, was employed for the molecular characterization of a gasoline sample in order to analyze the functional groups. The chemical composition of the gasoline sample was then expressed using six hydrocarbon functional groups, as follows: paraffinic groups (CH, CH2 and CH3), naphthenic CH-CH2 groups and aromatic C-CH groups. The obtained functional groups were then used to predict a number of fuel properties, including research octane number (RON), motor octane number (MON), derived cetane number (DCN), threshold sooting index (TSI) and yield sooting index (YSI).

Quantitative extraction and determination of trace elements by surfactant-free liquid-liquid microextraction from aviation and motor fuels

The occurrence of metals in crude oil and distillates may adversely affect oil processing, engines and turbines, plus contributing to environmental pollution, even at trace levels. Their detection is accordingly of the utmost relevance and several sample treatments have been developed for oily matrices, like sulfate ashing, microwave assisted digestion, dilution in organic solvents, and liquid–liquid (LL) extraction. The latter showed promising performances, but still suffer from several issues, like validation for a limited number of elements, lack of a clearly defined determined fraction (i.e., total concentration vs. dissolved fraction), unsatisfactory LODs, the need for matrix matched calibrations, and limited preconcentration factors. Here the development and validation of a liquid–liquid microextraction procedure without any surfactant is presented enabling the detection of the total concentration of 12 trace elements (V, Cr, Mn, Fe, Ni, Cu, Zn, Mo, Ag, Cd, Ba and Pb) in jet fuel and gasoline, and overcoming most of the mentioned drawbacks. The composition and volume of the extractant, plus the extraction time and temperature were systematically assessed by a multivariate approach, highlighting that the extractant volume is the most important factor controlling the extraction rate. Quantitative extraction of trace elements in jet fuel and race gasoline samples was achieved only under harsh conditions (aqua regia extractant, 50 °C, one hour, with an enrichment factor equal to 20) as successfully validated by spiking experiments and the standard sulfate ashing method ASTM D5708. Quantification was performed by both ICP-MS and Total Reflection X-Ray Fluorescence (TXRF), the latter fully exploiting the high enrichment factor achieved by the proposed LL microextraction method.

Evaluation of dried matrix spot sampling for total sulphur determination in automotive gasoline by high-resolution continuum source molecular absorption spectrometry and direct solid sample analysis

In this work, a method for total sulphur determination in automotive gasoline using dried matrix spot sampling is proposed. The method is based on the deposition of the sample on a cellulose-based filter paper and subsequent sulphur quantification via CS diatomic molecule using high-resolution continuum source graphite furnace molecular absorption spectrometry (HR-CS GF MAS). The sample deposition was carried out, along with the chemical modifier, on a 20-mm filter paper disc previously adapted into a polytetrafluoroethylene (PTFE) mould. The liquid phase was removed by heating the PTFE mould, and then the gasoline sample-embedded filter paper was punched in smaller discs (procedure A) or pulverised (procedure B) before the analyses. The mixture of Pd and Mg was used as chemical modifier to stabilise the sulphur compounds on the filter paper and on the graphite furnace. All the calibration curves constructed using seven different sulphur-containing compounds had a coefficient of determination higher than 0.995 and a linear range from 2 to 150 mg kg−1 S. By using the optimised conditions, the best characteristic mass, limits of detection and quantitation were 6 ng, 0.6 and 1.8 mg kg−1, respectively. The two sampling procedures (A and B) were evaluated for real samples, and procedure B was chosen since it markedly improved the precision. Using this procedure, satisfactory recovery values from 95 to 106% were obtained in the spike-recovery tests. In addition, the S concentrations for the certified reference materials were not statistically different from the certified values at 95% confidence level. Sulphur concentrations from 20 to 46 mg kg−1 were found in the six analysed gasoline samples, and these values were statistically assessed using a reference method (ASTM 5453). Spectral interference caused by MgF and MgCl diatomic molecules was observed and investigated.

Multi-output chemometrics model for gasoline compounding

Computational models may help to reduce research cost by predicting properties of alternative blends. Nowadays, most efforts focus on prediction of a few properties for sets of gasoline samples. However, there are no reports of models able for classification of gasoline samples with multiple output properties measured in real life refinery plants. In this work, Information Fusion (IF), Perturbation Theory (PT), and Machine Learning (ML) algorithm (IFPTML) was used to model real production data with >230,000 outcomes gathered from a petroleum refinery plant. IF-pre-processing phase assembled the working dataset with 44 physicochemical output properties vs. 574 input variables of 4 production lines distributed in 26 data blocks including 14 different streams and 23 operations carried out in the plant. PT-calculation phase quantifies the effect of perturbations (deviations) in all input variables using PT Operators. Last, in ML-analysis phase involved Linear Discriminant Analysis (LDA) and Artificial Neural Networks (ANN) models training. IFPTML-LDA model presented AUROC = 0.936 with overall Sensitivity Sn and Specificity Sp ≈ 84–91% for training and validation sets. In internal control experiment we obtained an IFPTML-FT-NIR model with similar Sn and Sp ≈ 86–97%, for >25,000 values of 16 properties measured FT-NIR technique; demonstrating the robustness of the algorithm to changes on the experimental techniques used. This model could be useful for the design of new alternatives blends (biofuels, refuse-derived fuels, etc.) with lower environmental impact.

Comparison of different processing approaches by SVM and RF on HS-MS eNose and NIR Spectrometry data for the discrimination of gasoline samples

In the quality control of flammable and combustible liquids, such as gasoline, both rapid analysis and automated data processing are of great importance from an economical viewpoint for the petroleum industry. The present work aims to evaluate the chemometric tools to be applied on the Headspace Mass Spectrometry (HS-MS eNose) and Near-Infrared Spectroscopy (NIRS) results to discriminate gasoline according to their Research Octane Number (RON). For this purpose, data from a total of 50 gasoline samples of two types of RON-95 and 98-analyzed by the two above-mentioned techniques were studied. The HS-MS eNose and NIRS data were combined with non-supervised exploratory techniques, such as Hierarchical Cluster Analysis (HCA), as well as other supervised classification techniques, namely Support Vector Machine (SVM) and Random Forest (RF). For supervised classification, the low-level data fusion was additionally applied to evaluate if the combined use of the data increases the scope of relevant information. The HCA results showed a clear clustering trend of the gasoline samples according to their RON with HS-MS eNose data. SVM in combination with 5-Fold Cross-Validation successfully classified 100% of the samples with the HS-MS eNose data set. The RF algorithm in combination with 5-Fold Cross-Validation achieved the best accuracy rate for the test set with the low-level data fusion system. Furthermore, it allowed us to identify the most important features that could define the differences between RON 95 and RON 98 gasoline. On the other hand, using the HS-MS eNose and NIRS low-level data fusion reached better results than those obtained using NIRS data individually, with accuracy rates of 100% in both SVM and RF performances with the test set. In general, the performance of the SVM and RF algorithms was found to be similar.