Engine oil quality deterioration estimation using an integrated sensory system

Engine oil degradation impacts the operation and performance of an internal combustion engine. As a result, it is crucial to compare the degradation level of engine oils, with a particular emphasis on kinematic viscosity, oxidation, nitration, and critical oil testing characteristics. Suppose engine oil changes too soon without first calculating its remaining usable life. In that case, it wastes already scarce resources and has an unfavourable environmental influence. Engine performance may suffer if oil is changed too late and is of poor quality. In the current research, oil testing is performed on vehicles due to maintenance to find the best moment for changing engine oil. Oil samples were taken randomly from cars that came to an authorized service facility for maintenance, comprising a wide range from the first to the fifth servicing. Oil samples were initially evaluated in a laboratory using a viscometer and Fourier transform-infrared spectroscopy in conformity with industry standards. The samples were then evaluated using an integrated sensor system developed and built by the authors. The determination coefficient R2 = 0.97 showed a significant positive association between major engine oil degradation characteristics such as kinematic viscosity, oxidation, and nitration at reasonable sensitivity. Approximately 8% of the total gathered samples were useable. The research identifies several oil degradation sensor systems that give low-cost, on-site convenient options for oil condition monitoring and predicting the remaining usable life of engine oil.


Introduction
Internal combustion engine operations and output are affected by engine oils and different operating conditions that affect the acceleration of oil degradation. 1Various parameter define the safe level of operation for engine oil.Above this safe level, it will affect the working and performance of the engine. 2The engine oil itself is a complex hydrocarbon compound. 3Along with the ageing process of the engine oil some contaminations are also added during the service period of the engine oil such as wear debris, glycol, soot, water, etc. 4 Based on the above facts the determination and quantification of the engine oil deterioration are really complex because it is also dependent on the driving conditions 5 and the engine type. 6As per guidelines provided by the vehicle manufacturers there are recommendations for the engine oil change cycle.But these recommendations are based on standard driving conditions.Considering that the travelling distance of vehicles is much less than a particular kilometre range (if proper driving conditions are followed) then there are chances that such engine oil may not deteriorate.It is a known fact that most of the crude oil required in India for manufacturing engine oil is imported, so there is a need for quantification of the deteriorated engine oil before replacement during a regular servicing schedule.Secondly, monitoring engine oil deterioration will provide important information about the possibility of engine failure. 7To quantify engine oil deterioration level various laboratory methods are available depending on parameters, such as viscosity, nitration, oxidation, sulfonation, total acid number (TAN), total base number (TBN), water and glycol content, or determination of antioxidant content. 8This provides the actual oil deterioration level, but most of these methods are destructive.Also, the costs associated with these methods are higher as compared to the cost involved in changing the engine oil (light-duty vehicles).
There can be offline condition monitoring or online condition monitoring of engine oil under use.For large stationary engines definitely, an online condition monitoring method is required.It is obvious that the cost of the online condition monitoring method is higher than the offline condition monitoring method.In some light-duty vehicles, the engine oil condition is determined approximately by some of the vehicle parameters such as the temperature of the engine, the duration of travel, the speed of the vehicle, etc. 9 There are some disadvantages of the offline laboratory analysis.Oil samples must be properly collected over specific intervals.Sample testing will be costlier because of precision laboratory equipment cost.This will consume more time in testing.Also, it involves sample handling time from the vehicle to the testing facility.This leads to the need for additional time to sample and analyse the results.
Infrared analysis (IR) is a powerful tool for the analysis of the degradation of engine oil. 10,11Parameters such as oxidation, nitration, etc., along with contaminants such as fuel, glycol, 12 soot, sulphate by-products, anti-wear components, etc., can be traced using the IR absorption spectrum 13,14 or atomic force microscopy. 157][18][19] Similarly, the oxidation level of deteriorated engine oil can be predicted by the ultravioletvisible (UV-Vis) micro-volume spectrometer instrument. 15Also, various latest handy instruments were developed for degradation analysis. 20The viscosity of the oil can physically increase because of oxidation. 21ll covalent chemical bonds in organic molecules absorb IR radiation at a characteristic wavelength, based on the type of antioxidants in the material, anti-wear components, hydroxyl, carbonyl compounds, sulfonate detergents, etc.According to a literature survey, absorption peaks are observed near wave numbers 970 and 860 cm −1 .These numbers have a relationship with absorbance and oxidation times.Each one of these oxidation times varied significantly compared to all other times. 22The Terahertz Time Domain Spectroscopy method demonstrated an oxidation time measurement. 23revious research showed the possibility of using colour features to analyse particles for automated wear debris classification. 24Published literature shows the use of an optical colour sensor with a blue LED as a light source and light-dependent resistor by infrared sensor setup 1 (LDR) or colour scale as per the ASTM D1500 standard for colour analysis. 25That displays a strong correlation to a change in voltage across LDR with the pH value and the viscosity of the oil. 26The RGB sensor output in three wavelengths viz.red, green and blue demonstrate a good correlation to viscosity, TAN, and Fe-particle concentration. 27Artificially degraded oils, output results are also compared with a viscometer, TAN and UV-Vis photospectrometry. 28revious studies have emphasized the use of a colour sensor to measure the deterioration in the TAN.The TAN and oxidation parameters increase resulting in the darkening of the colour of the oil. 29Also, a study of change in colour measurement for the L* a* b* system by colorimeterspectrophotometer CM-3500 shows a negative correlation between kilometres travelled. 30An opto-resistive quasi-digital sensor demonstrates measurement technique to quantify deterioration level. 31revious research has demonstrated that the concentration of contaminants (Pb, Cr, Fe, Ni) 32 and chemical and physical properties (such as TBN, TAN, and KV) in engine oils change over time during their service life. 33,34 oreover, the latest literature reports semi-quantitative tracking of the depletion of additives such as zinc-dialkyl-dithiophosphate (ZDDP). 35This monitoring also includes polymerized degradation products, soot, and wear particles with probable cause. 36To quantify engine oil deterioration various other sensors are also used such as capacitive, conductive, 37 inductive, optical, acoustic sensing, dielectric spectroscopy, permittivity and integrated sensors for measuring multiple oil properties. 38uartz tuning forks and extensional microresonators are used to demonstrate the measurements for viscosity and density. 39Higher-order model analysis is advantageous in the viscosity and density calculation from measured resonance parameters. 40arious existing research in the wider literature has been explored.Published literature includes a study conducted on a few civilian vehicles under typical urban traffic conditions. 41,42The majority of prior research has used either 'laboratory deteriorating engine oil' from a restricted sample number or the same vehicle engine oil during a set time of testing.According to a survey of the literature, no previous research has investigated an integrated sensors strategy for quantifying engine oil quality deterioration using the subsequent sensors.
This paper reports a novel method to address deterioration estimation by the collective effect of kinematic viscosity, oxidation and nitration.A large number of random samples and their analysis were used to use a novel method for the estimation of deterioration in terms of various parameters.This accomplishes the objective of bridging a research gap.The objective is to conduct experimentation on randomly obtained field samples of engine oil (more than 700) from various cars across a broad range of distances traversed.The major purpose of this research was to develop a low-cost integrated sensor configuration for quantifying engine oil degradation by comparing findings from different sensors.The developed novel configuration consists of a low-cost sensor array, which includes a low-cost IR sensor setup, a fluid property sensor (FPS), and a colour sensor.The study was a report of an investigation based on how certain physicochemical properties of engine oils change when they are used in real life.Studies related to such changes in physicochemical properties of selected engine oil parameters viz.degree of oxidation, KV and the degree of nitration, are presented in the 'Results and discussion' section.

Materials and methods
To prove the feasibility of the integrated sensors setup used for measuring deterioration in the oil in terms of oxidation, KV and nitration, samples of oil were gathered at random intervals from cars just after they had an oil change.SAE class 5W-30 engine oil samples (more than 700) were gathered from an authorized service station operating in the city of Pune, India.
With the main aim to identify the effect of deterioration in the engine oil, the following parameters are investigated in this research work: • Kinematic viscosity (ASTM D7042) was measured with a laboratory equipment viscometer (Make: Stabinger Viscometer model: 'SVM 3000').• Oxidation and nitration (ASTM E2412/D7624) were measured by an FT-IR spectrometer (make: 'Nicolet iS10' FT-IR spectrometer).
In the earlier published literature, 43 the experimental approach for measuring oxidation and nitration on FT-IR lab equipment is detailed.SVM 3000 Stabinger Viscometer with cylinder shape measures KV 0 .The authors analysed the OEM brochure and use instructions.
It uses a modified Couette concept with a fast-rotating outer tube and a slower-rotating measuring bob.A spinning magnet in the SVM 3000 creates a speed-dependent eddy current field.This device measures KV, dynamic viscosity (0.2-20 mPa•s), density (0.65 to 3 g/cm 3 ), and temperature (−56 °C to 105 °C).For an oil sample, all you need is a 2.5 ml measuring cell.This is needed to compute the viscosity index of degraded engine oil to assess crucial degradation characteristics.Using the above-mentioned laboratory equipment, degraded engine oil samples were analysed at Valvoline Cummins Pvt.Ltd, Ambernath, Thane.After testing, eight samples from the integrated sensors setup were rejected due to excessive degradation.This study analyses and tests the remaining oil samples.
According to the published literature, 10 engine oil samples are obtained for first, second, third, and higher servicing.The light-duty 4-wheeler service schedule suggests servicing after 10,000 km or 12 months, whichever came first 46.1% of the samples are from the first vehicle service, 37.2% from the second, and 16.7% from the third or later.Create a database of used engine oils.Next, we examine the suggested integrated sensors setup to measure these characteristics using a low-cost convenient system.

Integrated sensors setup details
The proposed integrated sensor setup to quantify engine oil deterioration has multiple sensors as shown in the schematic representation and the photograph of the actual setup is shown in Figure 1.An integrated setup block diagram displays encapsulated system functional modules.This system includes an FPS2000 sensor for measuring viscosity, density, DC, and T f .Next are two IR sensor transmitter-receiver pairs (IR sensors setup 1 ).A TCS3200 colour sensor measures changes in red.Also, included are a second IR sensor transmitter and receiver for measuring transmittance.A Robo LM35 sensor monitors ambient temperature throughout the experiment.This sensor is coupled to a CentIoT-MCP2515 and an Arduino CAN Bus Receiver TJA1050.All these sensors are linked to an Arduino Mega 2560 board through an Analogue-to-Digital Connector, where data is gathered and converted from hex to digital.This data is transmitted to the Raspberry Pi 4, where it is processed and displayed using GUI software.Because of the Raspberry Pi 4's connector layout, a keyboard and mouse may be attached for the easy code change.The next section describes each sensor's application.
The first is a Parker Hannifin Manufacturing (UK) Ltd FPS2000 FPS.Previous literature discusses the sensor's construction, operations, and workings. 4This sensor measures KV at the testing temperature of oil, dielectric constant, density, and dynamic viscosity.The dielectric constant gives further information on engine oil deterioration. 44Previous studies have emphasized that there is an increase in the dielectric constant with a longer distance travelled by the vehicle. 45The sensor's tip must be dipped in degraded oil.The sensing tip was totally submerged in a 10 ml oil sample in a glass beaker for the experiment.The second sensor set is an IR sensor transmitter and receiver pair mounted to a glass tube with Teflon support. 46The IR sensor pair is situated at a cross position to the glass tube so that the sensors can detect lighter engine oil and dark, deteriorating engine oil.In the final setup 1 , black Teflon supports prevent light from affecting the IR sensor's response.Large chamfers surrounding the tube promote new oil detection.IR transmitter and receiver distance must be tuned.For accurate detection, a larger-diameter funnel to avoid bubbles and a silicon tube for connecting to the glass tube are used.When oil passes the first pair of IR sensors, the timer begins, and when it reaches the other pair, it ends.Using these sensors, we can compute the oil flow time at room temperature by calculating KV.Readings are performed at 28 °C and varied room temperature to validate sensor set functioning.Details are in the paper's results and discussion.The IR pair measures the output in milliseconds to determine KV.If time decides the output, sensitivity analysis for this sensor type is not needed.
The third sensor set IR transmitter and IR receiver pair (IR sensor setup 2 ) is fitted on the top and bottom surfaces of the right side of a black colour plastic enclosure.The effect of the ambient light can be avoided by providing a separate enclosure for optical measurement that is the change in transmittance.The working principle with a schematic representation is discussed by researchers. 10Light from the IR transmitter is transmitted through the deteriorated engine oil sample towards the IR receiver.Some portion of the light is absorbed or has resistance to pass through the deteriorated engine oil because of the change in the physical and chemical properties.
Hence, the intensity of light is different at the IR receiver and is converted in corresponding resistance in terms of mv.The IR sensor setup 2 shows a sensitivity of 0.0365 for measurement of transmittance (mv).
The fourth suggested sensor is the RGB colour sensor TCS3200, with customizable colour light-to-frequency converters on a single monolithic CMOS integrated circuit.To the best of the authors' knowledge, RGB sensors have never been used to measure oxidation.The output is a square wave with a frequency proportional to light intensity (irradiance).The TCS3200 uses 8 × 8 photodiodes.Red, blue, green, and no filters are on 16 photodiodes.The datasheet shows a greater relative red responsivity between 600 and 800 nm.This colour sensor has black enclosures.A 1 mm thick oil film in a white plastic container is needed for the experiment.Oil samples are always positioned beneath the colour sensor for consistent results.The colour sensor has a sensitivity of 0.051 for a range from 570 mV to 400 mV, or until 50% deterioration level in terms of oxidation; if it has a sensitivity of −0.165; the oil sample is virtually at the end of its useable life and is almost constant in oxidation.
Figure 2 shows a quantifying model of the engine oil degradation of various parameters.Based on discussions of the methods and materials used, mathematical models that represent lubrication oil degradation due to oxidation, nitration and KV 0 that was induced in a laboratory in terms of KV are shown.Given that the experimental method has some inherent disadvantages such as a limited scope of degradation coverage, unavoidable test errors and longer periods of pre-installation training time.It is possible to use mathematical models to simulate the actual health status of the oil samples by making use of a sensor output that reflects the inputs of all sensors.Multiple devices were utilized to observe engine oil degradation in terms of oxidation, nitration, and alteration in the transmittance measurement.The next section contains a discussion of the results.

Results and discussions
The measurements for permitted limits for both oxidation (A/cm) and nitration (A/cm) are 25 (A/cm) and were given by engine oil producer Valvoline Cummins Private Ltd.Any reading over that necessitates the replacement of the engine oil (5W-30).The oxidation [A/(0.1 mm)] must be <0.4,that is, 40 A/cm, according to the published literature. 8Based on KV, the permitted limit for engine oil consumption is ±15% when compared to new engine oil.Based on this data, Table 1 displays the deterioration level scale for measuring engine oil degradation for parameters KV 0 , nitration, and oxidation.This scale will indicate if the engine oil under test is suitable for further RUL testing.Using this scale, % degradation levels relative to utilised (new) oil can be defined.
First, the experimental observations on multiple regression analysis for KV 0 measured at 40 °C by a Viscometer are compared with KV 1 at T f (please refer to Table 1).
This multiple regression analysis proves that the P-value of the experimental data is under the acceptable limit.R 2 = 0.93, the correlation for 95% confidence level shows a fairly good correlation between both methods of measurement for KV.
From the experimentation we get multiple regression equation (1) between the laboratory accurate method for KV 0 and KV 1 at T f : where X 1 is KV 1 measured at T f .Similar to the above method, simple regression analysis between the KV 0 measurement at 40 °C is compared with the KV 2 by calculation for the time needed for the oil to flow through IR sensors setup 1 at 28 °C, that is, at constant temperature in the laboratory (please refer to Table 2).
A good correlation between both methods of measurement is observed with R 2 = 0.93 and the P-value of the analysis shows that a correlation exists between both methods at a 95% confidence level.This also shows a linear correlation in the form of equation ( 2): where X 2 is the KV 2 at a constant temperature, that is, at 28 °C.
To validate the output from the IR sensors, setup 1 readings are taken at variable room temperatures.Also, during  the actual use of this integrated setup during on-site/field operation, measurements will be carried out at room temperature only.Hence, multiple regression analyses for these readings with respect to KV 0 measured at 40 °C are done.(Please refer to Table 3 for the results of the analyses).
It is similar to the earlier method with a higher correlation R 2 = 0.96 because the actual temperature factor measured by the temperature sensor of the system is nearer to the laboratory method.A slight variation in the temperature may be there during the constant temperature measurement at 28 °C.Equation ( 3) demonstrates the mathematical model to calculate the KV in terms of the KV 2 : where X 3 is KV 2 at T f in °C.
Sensor measurement validation for the analytically calculated value of the kinematic viscosity measured by the IR sensor for 28 °C or kinematic viscosity measured by the IR sensor for fluid temperature and kinematic viscosity at fluid temperature is shown in Figure 3.These analytically calculated values are to calculate KV at 40 °C plotted versus KV 0 measured at 40 °C.Also, results obtained by both sensors, that is, KV 2 at T f and KV 1 at T f versus KV at 40 °C by viscometer are presented in a graph in Figure 4.
Results obtained analytically using equations ( 1) to ( 3) are demonstrated graphically in Figure 3 which shows that results obtained by both sensors are comparable whereas results obtained by the IR sensor pair setup 1 show the highest correlation of R 2 = 0.96.Using these two methods for the integrated sensor setup, we can have  two versions.Version 1 will have only the low-cost IR sensor pair setup 1 for calculating the KV in the field trial.In version 2 , we can have both sensors installed as discussed in this research work.By using version 1 , the cost will be reduced, but there are some limitations of version 1 , such as the DC value for that sample will not be available.Using version 2 , the deteriorated engine oil sample, which has a higher level of soot or due to other deterioration parameters, can be identified with better accuracy.KV 1 will not be accurate at the higher deterioration level because it works on the tuning fork sensing method, where very same displacement amplitudes at the resonator surface with comparatively high frequency are observed.This shows an altered (e.g.non-Newtonian) measured viscosity behaviour when analysed to a current rotating viscometer evaluation.As a result, the sensor output has no correlation with a classical viscosity test in engine oil and contains high molecular weight viscosity modifier ingredients.This occurs as sensing probes an oil surface and will not find any effects induced by structural or complex interactions with related typical lengths equal to or higher than the acoustic wave's depth of penetration in the oil. 47sed engine oil samples that were gathered from the first servicing and tested by the viscometer showed that 244 samples had deteriorated, that is, 74.4% of samples had deteriorated out of 328 samples.In total, 184 samples from the second servicing had deteriorated out of 264 samples, that is, 69.7% of samples had deteriorated.From the third and later servicing schedule 84 samples were found to have deteriorated for KV out of 112 samples, that is, 75%.This shows that considering all engine oil samples 72.7% of samples had deteriorated and had to be replaced.However, the focus of this experiment is on engine oils which have not deteriorated and would still be replaced following standard procedure.
The FPS sensor's measurement information is insufficient to cover far-end samples.Hence, IR sensors setup 1 will rectify this.Combining FPS and IR sensor setup1 measurements yields good results, even if the improvement is negligible.
The second experimental findings were for the measurement of oxidation (A/cm) with respect to the red colour frequency output measured by the colour sensor (C R ).During experimentation, engine oils which had deteriorated equally to or above the oxidation value of 35 A/cm were removed from the analysis because these deteriorated oil samples could have affected our conclusion.We have considered a range of 40% above the deterioration level of 25 A/cm.Figure 5 shows a plot for experimentation results displaying the change in oxidation versus variation in C R .
Simple regression analysis between oxidation (A/cm) measured in the laboratory versus alteration in C R for degraded oil samples is carried out.Referring to the experimental investigation plot shown in Figure 5, the linear correlation coefficient shows R 2 = 0.86.Taking that into consideration and also the P-value and other statistical results of the regression analysis, a valid relationship between the parameters of the experimentation is observed.But looking at the nature of the data quadratic fitted model, the tread line shows R 2 = 0.92.Hence, we can use the quadratic fitted model tread line to derive a mathematical model between the methods under experimentation which is shown in equation ( 4): where X 4 is the C R at room temperature.The value of X 4 can be used to calculate the oxidation value by an analytical method.
Used engine oil samples collected from the first servicing tested by FT-IR spectroscopy showed that eight samples had deteriorated, that is, 2.4% of samples had deteriorated out of 328 samples.Sixteen samples had deteriorated out of 264 samples, that is, 6% of samples had deteriorated from the second servicing schedule.From the third and later servicing schedule, four samples had deteriorated for oxidation out of 112 samples, that is, 3.57%.This shows that considering all engine oil samples for KV and oxidation 72.7% of  samples had deteriorated and had to be replaced.However, we need to focus on engine oils which had not deteriorated but still had to be replaced following standard procedure.The third experimental findings are for the measurement of nitration (A/cm) with respect to transmittance (mv) value measured as the output of the IR sensor setup 2 .During experimentation, engine oils which had deteriorated to equal or above the nitration value of 45 A/cm were removed from the analysis because these had deteriorated oil samples which could have affected our conclusion.We have considered a range of 50% above the deterioration level of 25 A/ cm. Figure 6 shows the results plot of the experimentation undertaken for the change in nitration to variation in transmittance (mv) output by the IR sensor setup 2 .
Simple regression analysis between nitration (A/cm) measured in the laboratory versus T IR2 for a degraded sample of oil is carried out.
Referring to experimental investigation plots and regression analysis between nitration and T IR2 and the linear correlation between data a good correlation coefficient of R 2 = 0.93 is observed, where the P-value of the analysis is below the recommended value of 0.01.Hence, with more than a 95% confidence level we can have a linear correlation trade line to derive a mathematical model between the two parameters and Figure 6, which is shown in equation ( 5): where X 5 is the T IR2 (mv) value at room temperature.The value of X 5 is used to calculate the nitration value by an analytical method.
The following discussion for the nitration parameter related to the said engine oil samples collected from the first servicing tested by FT-IR spectroscopy showed that 100 samples had deteriorated, that is, 30.5% of samples had deteriorated out of 328 samples.Eighty samples had deteriorated out of 264 samples, that is, 30.3% of samples deteriorated from the second servicing schedule.From the third and later servicing schedule, 40 samples had deteriorated for nitration out of 112 samples, that is, 35.7%.
This shows that considering all engine oil samples for KV, oxidation and nitration, 93.18% of samples had deteriorated and all the collected samples were from vehicles that had travelled a distance of 10,000 km or more than 12 months had elapsed since their previous servicing.To the best of the authors' knowledge, it is for the first time that such a large number of oil samples have been used in the experimentation.Also, during field trials usually, information about the oil kilometre at that servicing schedule can be easily available as compared to the number of servicing done before the current servicing schedule.
Comparing our findings to those of prior research, it must be noted that our suggested integrated sensor setup gives extensive information on KV, oxidation, and nitration.Using multiple regression analyses of data collected from all the sensors and input from users as to the oil kilometres to quantify the engine oil deterioration in terms of KV is reported in the next part of the discussion.The multiple regression equation is shown by equation ( 6) and analysis details are provided in Table 2: where = T IR2 and D = DC.Equation ( 6) and the statistical analysis together, confirm that by using equation (6) we can calculate the engine oil KV 0 at 40 °C similar to a viscometer, a precision and accurate instrument used in the laboratory to quantify engine oil deterioration by using a mathematical model technique.This delivers significantly better results due to the integrated sensor setup approach for experimentation.A large number of existing studies in broader literature have examined and reported that KV is an important engine oil parameter because if there is deterioration beyond the specified limit as stated by the standards the friction between engine components will definitely increase which will, in turn, affect its performance.Now using equation ( 6) to quantify engine oil deterioration all the samples which are found to be deteriorated by laboratory methods are exactly all the samples that are shown to be deteriorated by analytical calculations.
Experimental measurements are usually unreliable.Experiments investigate several inputs.These parameters are monitored using electronic devices.An integrated sensory system uses user-entered oil kilometres, ambient temperature assessed by the LM35 sensor, Fluid Property Sensor FPS2000 sub-sensor configuration, TCS3200 colour sensor system, and an IR sensor setup 1 for study.According to the calculations (analysing measures of uncertainty), the total error in measuring accuracy and rating engine oil degradation ranged from ± 0.5% to ± 2.1%.Overall measurement uncertainty is ± 3 to ± 8%.Given the application, this remains within an acceptable range.
Along with that, all the samples which showed a 99% deterioration level in laboratory experimentation are almost found to have deteriorated to the same level by the integrated sensor setup.Therefore, the output for the integrated sensors setup using equation ( 6) needs to be verified on a new scale which is provided in Table 3.
The measurement of the engine oil deterioration using the integrated sensory setup can provide procedures for handling data along with suggestions of logic provided by the information displayed in rows 2 and 3 of Table 3.Having determined how much the engine oil has deteriorated, further, it will be useful to figure out the remaining useful life (RUL).

Conclusions
The research reported in this paper fulfilled the identified research gap.An experimental investigation has been carried out with the randomly collected field samples of engine oil (5W-30).Quantification of oil degradation in terms of changes in oxidation and nitration values is obtained.It is evident from the experimental investigation that, an integrated low-cost sensory system could capture required information by IR sensor setup 2 and colour sensor with acceptable sensitivity (0.0365 and 0.051, respectively).
This result help to conclude that: • An integrated sensory system identified a better correlation for kinematic viscosity (R 2 = 0.96); oxidation (R 2 = 0.92) and nitration (R 2 = 0.93) could help us to quantify the quality of engine oil deterioration.• Based on statistically evident multiple regression analysis, it can be effectively used for predicting the RUL of the engine oil by determination coefficient R 2 = 0.97 with an acceptable accuracy level of 97% and root mean squared error of 2.619.
This study may be extended to measure the remaining usable life in kilometres, regardless of the vehicles' servicing schedule.Similar systems using additional sensors that measure water, glycol, and soot may more accurately estimate engine oil degradation.In summary, as per the objective of the research, the proposed integrated sensory system potentially offers a low-cost solution for offline estimation of engine oil quality deterioration.

Figure 1 .
Figure 1.Integrated setup block diagram and actual photo.

Figure 2 .
Figure 2. Model for the quantification of engine oil deterioration with respect to viscosity, oxidation, nitration, and indirect sensor-based monitoring techniques.

Figure 3 .
Figure 3. Analytical output comparison for KV 1 at T f and KV 2 at 28 °C or T f versus KV 0 at 40 °C.

Figure 4 .
Figure 4. KV 1 and KV 2 at T f versus KV 0 at 40 °C.Figure 5. Oxidation (A/cm) versus alteration in C R for degraded oil samples.

Figure 5 .
Figure 4. KV 1 and KV 2 at T f versus KV 0 at 40 °C.Figure 5. Oxidation (A/cm) versus alteration in C R for degraded oil samples.

Table 1 .
Scale to quantify engine oil deterioration in terms of KV 0 and nitration or oxidation for SAE 5W-30.

Table 2 .
Regression analysis between KV 0 versus O km , T f , KV 1 , KV 2 , C R , T IR2 and DC.

Table 3 .
Modified scale to quantify engine oil deterioration in terms of KV at 40°C considering the output of all sensors for SAE 5W-30.