WO2015142283A1

WO2015142283A1 - A method and device for determining the quality of edible oil

Info

Publication number: WO2015142283A1
Application number: PCT/SG2015/000080
Authority: WO
Inventors: Weibiao Zhou; Geeta BANSAL
Original assignee: National University Of Singapore
Priority date: 2014-03-18
Filing date: 2015-03-13
Publication date: 2015-09-24
Also published as: SG11201607651WA; CN106662561A; CN106662561B

Abstract

A method of determining the quality of edible oil is disclosed. The method comprises determining a plurality of parameters of the edible oil at 502, the plurality of parameters including dielectric constant and viscosity, using the plurality of parameters as input variables of a mathematical model for predicting values of both the total polar compounds (TPC) and triglyceride polymeric content (PTG) of the edible oil at 504, comparing the predicted TPC and PTG values with corresponding predetermined thresholds at 506, and determining the quality of the edible oil based on the comparison at 508. A corresponding device is also disclosed.

Description

A Method and Device for Determining the Quality of Edible Oil Field & Background

The present invention relates to a method and device for determining the quality of edible oil.

For deep frying of food, the edible oils used are exposed to oxygen and moisture from the food at high temperatures of around 180°C. As a result, the oils undergo numerous chemical reactions (e.g_; polymerization, and oxidation) and consequently generate a significant number of compounds that alter the quality of the original oils, resulting in oil deterioration over time. Specifically, oil deterioration is generally followed by changes in colour of the used oil, free fatty acid level, an increase in trans-fats, polycyclic aromatic hydrocarbons (PAHS) and the polarity of the oil. The oil deterioration issue is further exacerbated under commercial setting applications, since the oils are typically re-Used many times to save costs before discarding and thus may pose public health hazards. As the quality of food prepared in the oils is directly related to the quality of the oils, it is important to ascertain the extent of deterioration of the oils through degradation indicators to mitigate the health hazard of using those oils. There is however no easy and straightforward method for determining the indices to ascertain when the oil has deteriorated to an extent to necessitate discarding.

As widely appreciated, the oils used for cooking must be discarded after a period of use. Conventionally, the decision on when to discard the oils is based on visual inspection of the oil quality, or through the experience of a cook who decides when to change the oil based on colour, odour, excessive foaming and smoking, and/or simply by tasting the fried food products. Due to the subjective nature of these methods, they are however not reliable, and alternative methods that provide quantitative information on the degradation indices are needed to ensure the health and safety of public consumers of the fried food products. Currently, the various chemical and physical parameters described by the American Oil Chemists Society (AOCS) are used to evaluate oil quality in laboratory settings. Of those parameters, the German Society of Fats and Oils (DGF) further recommends focusing on two important quality parameters, total polar compounds (TPC) and polymeric triglyceride content (PTG) for such oil quality evaluation. However, in regular operation of food establishments, determination of these two quality parameters to measure oil quality through frequent analytical tests is difficult and challenging due to the time and costs involved. Kits that enable rapid testing are instead desired for providing practical, quick and easy means to monitor the deterioration of oil quality.

Unfortunately, there very few of such testing kits on the market. Further, of those available commercially, a major drawback of the testing kits is that they monitor only one aspect of the oil quality, such as free fatty acids (FFA), oxidised compounds, or TPC, which has been shown to be inadequate according to investigations. Moreover, it was also demonstrated that none of those testing kits was able to provide, on a statistical level, comparable results as those obtained via the analytical methods for commonly used frying oils. This can be seen from the graph 100 and chart 200 of Figures 1 and 2, which respectively depict comparison of results obtained from two different approaches (one being a testing method and the other being a testing kit), and the corresponding results with formal analytical methods. There have also been efforts to devise better methodologies for judging the oil quality based on measurements of multiple degradation indices, but without much success either.

One object of the present invention is therefore to address at least one of the problems of the prior art and/or to provide a choice that is useful in the art.

Summary

According to a 1^st aspect of the invention, there is provided a method of determining the quality of edible oil. The method comprises determining a plurality of parameters of the edible oil, the plurality of parameters including dielectric constant and viscosity, using the plurality of parameters as input variables of a mathematical model for predicting values of both the total polar compounds (TPC) and triglyceride polymeric content (PTG) of the edible oil, comparing the predicted TPC and PTG values with corresponding predetermined thresholds, and determining the quality of the edible oil based on the comparison. The proposed method advantageously enables rapid testing of the quality of repeatedly re-used edible oils during and after repeated deep-frying operations and also allow the testing to be carried out in a convenient and practical manner, for determining if those oils are over-used for the benefit of food safety purposes. More specifically, the proposed method provides prediction of both the values of two important quality indicators of the edible oils (being the TPC and PTG) by utilising rapid measurements of certain input variables, which enables determination of the extent of chemical deterioration the edible oils have undergone, and consequently, if the oils are thus still fit for use. Furthermore, the proposed method is able to reliably provide accurate predicted values of both the TPC and PTG.

Preferably, the mathematical model may include a nonlinear mathematical model. More preferably, the nonlinear mathematical model for predicting the value of the TPC may be expressed as an equation: TPC = -12.353 - 0.0199DC² + 1.591DC + 0.00519vis² - 0.0328vis, where DC is the dielectric constant and vis is the viscosity. Further, the equation may be characterised with a root-mean- square deviation (RMSE) value of 2.01 , and with a R² value of 0.93, where R is the Pearson product-moment correlation coefficient.

Yet preferably, the nonlinear mathematical model for predicting the value of the PTG may be expressed as an equation: PTG = -9.936 + 0.00608vis² - 0.0385vis - 0.O0470DC² + 0.376DC, where DC is the dielectric constant and vis is the viscosity. Specifically, the equation may be characterised with a root-mean- square deviation (RMSE) value of 1 .50, and with a R² value of 0.92, where R is the Pearson product-moment correlation coefficient.

Moreover, determining the dielectric constant may include measuring the dielectric constant preferably at a temperature of about 80°C, while determining the viscosity may include measuring the viscosity preferably at a temperature of about 40°C. Additionally, the plurality of parameters may further include lightness, redness, blueness, absorbance and refractive index.

Preferably, determining the absorbance may include measuring the absorbance at a wavelength of 450 nm or 490 nm. Yet preferably, using the plurality of parameters as input variables of a mathematical model for predicting the values of both the TPC and PTG may further include computing the values of the TPC and PTG in parallel. Further preferably, the method may also further comprise generating an alert signal if the predicted TPC value or PTG value exceeds the corresponding predetermined threshold. More preferably, the method may further comprise providing a visual indicator corresponding to the generated alert signal.

Further preferably, the nonlinear mathematical model for predicting the value of the PTG may be expressed as an equation: PTG = -73.208 + 0.00692vis² - 0.0413vis + 33.317RI² + 15.473RI, where Rl is the refractive index, and vis is the viscosity. The equation may be characterised with a root-rhean-square deviation (RMSE) value of 1.44, and with a R² value of 0.94, where R is the correlation coefficient.

According to a 2^nd aspect of the invention, there is provided a device for determining the quality of edible oil. The device comprises a plurality of probes adapted to determine a plurality of parameters of the edible oil, the plurality of parameters including dielectric constant and viscosity, and a data processor adapted to: use the plurality of parameters as input variables of a mathematical model for predicting values of both the total polar compounds (TPC) and triglyceride polymeric content (PTG) of the edible oil, compare the predicted TPC and PTG values with corresponding predetermined thresholds, and determine the quality of the edible oil based on the comparison.

Preferably, the data processor may be further adapted to generate an alert signal if the predicted TPC value or PTG value exceeds the corresponding predetermined threshold. Further preferably, the device may also comprise a display screen for providing a visual indicator corresponding to the generated alert signal. The mathematical model may include a nonlinear mathematical model. Further, the nonlinear mathematical model used by the data processor for predicting the value of the TPC may preferably be expressed as an equation: TPC = -12.353 - 0.0199DC² + 1.591DC + 0.00519vis² - 0.0328vis, where DC is the dielectric constant and vis is the viscosity. Also, the nonlinear mathematical model used by the data processor for predicting the value of the PTG may be expressed as an equation: PTG = -9.936 + 0.00608vis² - 0.0385vis - 0.00470DC² + 0.376DC, where DC is the dielectric constant and vis is the viscosity. Alternatively, the nonlinear mathematical model used by the data processor for predicting the value of the PTG may also be expressed as an equation: PTG = -73.208 + 0.00692vis² - 0.0413vis + 33.317 Ri² + 15.473RI, where Rl is the refractive index, and vis is the viscosity. Preferably, one of the plurality of probes may be configured to measure the dielectric constant at a temperature of about 80°C, and also further, one of the plurality of probes may be configured to measure the viscosity at a temperature of about 40°C. In other words, values of the dielectric constant and the viscosity are may be obtained at different. temperatures.

Preferably, the plurality of probes may include at least two such probes. Specifically, one of the at least two such probes may be configured to measure both the dielectric constant and the temperature. Yet preferably, the device may further comprise an input panel for reconfiguring values of the predetermined thresholds. It should be apparent that features relating to one aspect of the invention may also be applicable to the other aspects of the invention.

These and other aspects of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter.

Brief Description of the Drawings

Embodiments of the invention are disclosed hereinafter with reference to the accompanying drawings, in which:

Figure 1 is a graph showing comparison of results obtained from two different approaches for measuring the quality of edible oil (e.g. sunflower oil), used for frying chicken nuggets according to the prior art, wherein the first approach is a testing method of AOCS %FFA and the second approach is a testing kit known as FASafe™ %FFA, wherein %FFA represents the percentage of free fatty acids; Figure 2 is a bar chart depicting the discrepancy between the experimental scores obtained by a commercial testing kit, Oxifrit-Test^®, and the corresponding expected scores obtained by formal analytical methods, for measuring the quality of palm oil used for frying French fries, according to the prior art;

Figure 3 is a scatter plot for validating an equation of a nonlinear mathematical model derived for predicting the total polar compounds (TPC) quality indicator of edible oil, according to an embodiment of the invention;

Figure 4 is a scatter plot for validating an equation of the nonlinear mathematical model derived for predicting the polymeric triglyceride content (PTG) quality indicator of edible oil, according to the same embodiment of Figure 3;

Figure 5 is a flow diagram depicting a method for determining the quality of the edible oil according to the embodiment;

Figure 6 is a schematic diagram of a device that incorporates the method of Figure 5;

Figure 7 is a bar chart comparing experimentally obtained TPC values against corresponding TPC values sampled from edible oils, obtained from random food establishments, using the device of Figure 6; and

Figure 8 is a bar chart comparing experimentally obtained PTG values against corresponding PTG values sample from edible oils, obtained from the random food establishments, using the device of Figure 6.

Detailed Description of Preferred Embodiments

According to a first embodiment of the invention, a method for rapid testing of the quality of edible oils (e.g. frying or cooking) based on two quality indicators that facilitate determination of oil quality is disclosed. The two quality indicators are the total polar compounds (TPC) and triglyceride polymeric content (PTG) of the specific edible oil being monitored, each of which is further a function of other input variables such as viscosity, dielectric constant, optical properties and etc.

It is importantly to be highlighted that the method, according to the first embodiment, is guided and based upon the development of nonlinear mathematical models specifically of the linear-plus-square type, by employing the aforementioned input variables that are able to cumulatively provide accurate predictions for the two quality indicators, TPC and PTG. In particular, the mathematical models are developed by way of stepwise regression techniques. The respective mathematical models developed for TPC and PTG are also independent of each other, in which each mathematical model incorporates eight input variables comprising lightness ("L*"), redness ("a*"), blueness ("b*"), absorbance (i.e. spectral density) at 450 nm ("ab₄₅o"), absorbance at 490 nm ("ab₄₉₀"), viscosity at about 40°C ("vis"), dielectric constant at about 80°C ("DC"), and refractive index ("Rl"). It is to be noted that the variables of lightness ("L*"), redness ("a*") and blueness ("b*") are measured using a spectrophotometer and expressed as the "L*a*b*" colour system. On the other hand, the absorbance is measured using a UV-Vis spectrophotometer over a spectrum of 400-700 nm. Further, the refractive index and viscosity are respectively measured using a refractometer and a viscometer. It will also be understood that the temperatures for measuring the viscosity and dielectric constant can also be performed . at other temperatures than those aforementioned, and accordingly the corresponding mathematical models adopted will thus also be different, as to be appreciated. Specifically, the relationship among these eight input variables is determined by computing the corresponding variance inflation factor ("VIF"). In this instance, the criterion of "VIF < 10" is used. VIF quantifies the severity of multicollinearity in an ordinary least squares regression analysis; VIF provides an index to measure how much the variance of an estimated regression coefficient is increased due to collinearity.

In developing the mathematical models, a commercial testing kit known as "Testo 265" (devised for measuring the TPC) is adopted for establishing a relationship between the DC and the TPC via a correlation model, which is developed as a quadratic model in this instance. Specifically, during the process of establishing the relationship, the TPC values and corresponding DC values of any edible oil samples are first respectively measured using the "Testo 265" testing kit, and another commercial laboratory instrument^ "DEA 2970". As known in the art, the "DEA 2970" instrument is configured to measure the dielectric properties of a material as a function of time, temperature and frequency. Mathematically modelling, using the quadratic correlation model, was then consequently carried out to analytically establish the relationship between those Testo-265-measured TPC values and corresponding DC values. As for the remaining seven input variables, experimental results obtained via analytical methods, are used for establishing the relationships between the TPC values and the seven input variables, as explained in the preceding paragraph. As afore described, linear-plus-square type mathematical models are developed for predicting the TPC and PTG quality indicators. With reference to the TPC prediction, it was determined, from evaluation of the various relationships as established between the TPC and respective input variables (as described in the preceding paragraph), that a best-fit mathematical model for quantification of the TPC quality indicator is one which incorporates at least the two input variables of the DC and viscosity. As set out below, the mathematical model as determined for predicting the TPC quality indicator is expressed as a second degree polynomial equation (1 ): TPC = -12.353 - 0.0199DC² + 1.591DC + 0.00519vis² - 0.0328vis (1 )

Equation (1 ) is characterised with a relatively low root-mean-square deviation (RMSE) value of "2.01 ", and with a high R² value of "0.93". It is to be appreciated that the term "R" is the correlation coefficient, while the term "R²" is the determination coefficient, and both are commonly used in statistical linear regression analysis to give an indication of how well a first term correlates to a second term. If the value of "R" is equal to it indicates there is perfect positive correlation between the first and second terms, but if the value is equal to it conversely indicates perfect negative correlation between the first and second terms, whilst- a value of "0" implies that there is no linear correlation between the first and second terms. Referring to equation (1 ), it is additionally highlighted that no multicollinearity issue was detected between the DC and square of viscosity input variables as defined in the equation. In other words, in statistical terms, this means that the DC and viscosity input variables are not highly correlated with respect to one another.

On the other hand, for the quantification of the PTG quality indicator, it was determined that a best-fit mathematical model that also does not suffer from the multicollinearity problem, was similarly one that incorporates the input variables of the DC and viscosity just like equation (1 ), and is expressed as equation (2) below:

PTG = -9.936 + 0.00608vis² - 0.0385vis - 0.00470DC² + 0.376DC (2)

Equation (2) is characterised with a RMSE value of "1.50" and a high R² value of "0.92", which is comparable to the corresponding properties of equation (1 ). Further, the performance validation results for equations (1 ) and (2) are respectively shown in the scatter plots 300, 400 of Figures 3 and 4.

Now with reference to Figure 5, which depicts a flow diagram 500 of the method, the quality of edible oil can be ascertained by first measuring the corresponding DC and viscosity values of the edible oil at step 502. Subsequently at step 504, the measured DC and viscosity values are input into equations (1 ) and (2) to respectively predict the TPC and PTG values of the edible oil. Then at a next step 506, the predicted TPC and PTG values are compared against corresponding predetermined thresholds, which are based upon food safety standards set by the relevant health authorities of a country. It will be apparent to skilled persons that the food safety standards vary between different countries. At last step 508, the quality of the edible oil is determined based on the comparison. More specifically, if any of the predetermined thresholds is exceeded, it means that the edible oil is no longer safe for consumption and should be discarded. Based upon the exceeding of any thresholds, an alert signal is generated and a visual indicator (such as a red indicator light) corresponding to the alert signal is raised to notify users. However, if none of the predetermined thresholds is exceeded, then the edible oil can still be used safely. A visual indicator (such as a green indicator light) is also raised to notify users in such an instance if the edible oil is still considered usable. The aforementioned mathematical models (i.e. expressed as equations (1 ) and (2)) derived for quantification of the TPC and PTG quality indicators, based on the method of the first embodiment, are then subsequently utilised for developing a new testing kit, to be used for rapid testing of the quality of edible oils. The new testing kit implemented in the form of a device 600 is shown in the schematic diagram of Figure 6. Specifically, the proposed device 600 comprises at least three probes 602a, 602b, 602c, an input panel 603, a display screen 604, and a data processor 606, which is implemented as an IC circuit that averages about the size of between one or two standard credit cards. The at least three probes 602a, 602b, 602c are configured for measuring the viscosity, dielectric constant and temperature respectively. Particularly, the at least three probes 602a, 602b, 602c are configured based on known operating principles/devices (e.g. capacitance-based for dielectric constant measurement, vibration-attenuation-based for viscosity measurement, and resistance temperature detectors for temperature detection) for performing the related measurements. It will be understood that any suitable off-the-shelf probes can be appropriately adopted for use as the at least three probes 602a, 602b, 602c. It is also further highlighted that in this instance, one of the probes 602a for measuring the viscosity is a portable ViSmart™ Sensor (which is available commercially). It will be appreciated that the display screen 604 and data processor 606 are properly housed within a casing 608 of the proposed device 600 for easy handling during operation. A custom developed graphic user interface (GUI) was then adopted as an operating interface for the proposed device 600. To determine the quality of any edible oil, a user (not shown) operating the proposed device 600, first inserts the at least three probes 602a, 602b, 602c into the edible oil and waits for the device 600 to measure the DC and viscosity of the edible oil being tested. Thereafter, once the DC and viscosity are measured, the data processor 606 computes the predicted values for the TPC and PTG of the edible oil, and compares them with predetermined thresholds stored in the device 600. The predetermined thresholds are defined according to food safety standards set by the relevant health authorities of a country, and thus it will be appreciated that the predetermined thresholds stored in the device 600 are configurable by the user via the input panel 603, depending on the food safety standards of which specific country the user wishes to follow for ascertaining the safety of edible oils for consumption. If the comparison reveals that either the TPC or PTG predicted value exceeds the corresponding predetermined threshold, an alert signal is generated by the data processor 606, which is then indicated as a corresponding warning red light display through the display screen 604 of the proposed device 600. The warning red light display informs the user that the quality of the edible oil is considered to be unacceptable for further usage, and needs to be discarded. Additionally, the TPC and PTG predicted values as computed are also shown on the display screen 604 to the user. Conversely, if the comparison reveals that none of the predetermined thresholds are exceeded by the PTC or PTG predicted value, a pass signal is then generated by the data processor 606, which is shown as a corresponding green light display on the display screen 604 of the device 600 notifying the user that the edible oil is still safe for food preparation. To assess the accuracy of results generated by the proposed device 600, the device 600 was subjected to some validation tests using various sets of experimental data, as well as being tested using oil samples collected from random commercial food establishments. With reference to the results of the validation tests as performed, the bar charts 700, 800 of Figures 7 and 8 showed that the proposed device 600 managetl to achieve fairly respectable performance. More specifically, the proposed device 600 performed particularly well for those oil samples with TPC and/or PTG values close to or exceeding food safety limits recommended by food health authorities, where it will be appreciated by skilled persons that these are situations which the accuracy of any testing kit is important for enabling determination of whether the edible oil, having being used for a period of time, should be discarded or still fit for use.

Further embodiments of the invention will be described hereinafter. For the sake of brevity, description of like elements, functionalities and operations that are common between the embodiments are not repeated; reference will instead be made to similar parts of the relevant embodiment(s).

According to a second embodiment, a different mathematical model was used for predicting the PTG quality indicator, whilst the mathematical model for predicting the TPC quality indicator is retained as described in the first embodiment, according to equation (1 ). In this instance, the mathematical model for quantifying the PTG instead utilises the two different input variables of the refractive index and viscosity, as opposed to the DC and viscosity. Specifically, the mathematical model for predicting the PTG quality indicator is expressed as a second degree polynomial equation (3): PTG = -73.208 + 0.00692vis² - 0.0413vis + 33.317RI² + 15.473RI (3)

It is to be noted that equation (3) is characterised with a low RMSE value of "1.44", and a high R² value of "0.94". Further, the proposed device 600, as described in the first embodiment, can then be modified accordingly to use the mathematical models expressed as equations (1 ) and (3) for rapid testing of the quality of edible oils. In summary, the proposed method solves the problem of enabling rapid testing of the quality of repeatedly re-used edible oils to be carried out in a convenient and practical manner, so that it can be determined easily whether those edible oils should be discarded. More specifically, the proposed method provides prediction of the values of two important quality indicators of the edible oil (i.e. TPC and PTG) by utilising rapid measurements of certain input variables (e.g. optical, rheological and dielectric properties of the oil) in conjunction with the aid of mathematical models, which consequently enables determination of the extent of chemical deterioration the edible oil has undergone, and hence if the edible oil is still fit for use. As such, the proposed method can be deployed under restaurants, food outlets and fast food outlets settings, where edible oils are reused repeatedly for frying/cooking food and decisions to discard those edible oils are conventionally taken based on subjective indicators such as the state of the oil colour, foaming conditions, or based on the results obtained through conventional testing kits, which have however been demonstrated by prior investigations to be inadequate and unreliable. It is also to be highlighted that the proposed method is also one of the first rapid testing method able to collectively provide accurate predicted values of both the TPC and PTG.

Further, it is also to be appreciated that the proposed method is developed for testing edible oils which have gone through deep frying operations. During deep frying, certain chemical changes occur due to the heat and produce new compounds in the edible oils, and thus the proposed method is formulated to test for occurrence of those new compounds in the edible oils to determine their usable quality. For example, if the edible oils are used only for stir-frying, those new compounds may not be present in an appreciable amount and therefore, the proposed method determines that the quality of the edible oil is still sufficiently satisfactory for safe usage and consumption.

The described embodiments should not however be construed as limitative. For example, the data processor 606 of the proposed device 600 may alternatively be realised in the form of a mini-computer external to the device 600 or any other suitable means of data processing device may also be used. Further, other types of mathematical models, including non-linear models, linear models, full quadratic models and the like, may also be used for predicting the TPC and PTG quality indicators, as understood by skilled persons. It will also be appreciated that other suitable mathematical models used for predicting the TPC and PTG quality indicators may also incorporate other input variables such as lightness ("L*"), redness ("a*"), blueness ("b*"), absorbance, and refractive index ("Rl"), in addition to the two input variables of the DC and viscosity. With reference to comparison of the predicted TPC/PTG values with the corresponding predetermined thresholds, instead of showing the warning red light/green light on the display screen 604 of the proposed device 600 when the thresholds are exceeded/not exceeded, other forms of indicators may be adopted such as displaying a corresponding suitable message, using audible signals or the like, as understood by skilled persons. Additionally, the proposed device 600 may also be configured to locally store data relating to multiple sample testing results that have been performed (i.e. historic data storage), and the stored data may easily be downloadable to external computers for analysis via a connection port provided on the proposed device 600. Furthermore, the stored data may also be displayable as a simple plot on the display screen 604 of the proposed device 600 to users for easy on-the-spot referencing when testing for the quality of edible oil.

In respect of the proposed device 600, instead of being configured with the at least three probes 602a, 602b, 602c, it may be adapted to work with at least two probes, in which a DC probe that is also capable of measuring temperature is used (i.e. the DC probe measures both DC and temperature simultaneously) as one of the two probes. Indeed, the proposed DC probe is understood to be equivalent to afore described two probes 602b, 602c that instead measure DC and temperature separately. It is to be noted that in such a configuration, the proposed DC probe is formed from a type of special metal that is typically used in resistance temperature detectors (RTDs) for temperature measurement (although the design and geometry adopted for the proposed DC probe may be different, but the variations possible will nonetheless be understood by skilled persons). Also in this instance, the proposed DC probe replaces the "Testo 265" testing kit, which is afore described in the first embodiment as being used for establishing the relationship between the DC and the TPC via a correlation model; in other words, the DC value is to be directly measured using the proposed DC probe, and there is no necessity to establish the relationship between the DC and the TPC anymore.

In addition, the proposed device 600 may not be configured with a related probe for measuring the DC values; instead, the DC values are indirectly measured using the "Testo 265" testing kit through use of a conversion model based on the measured output of the "Testo 265" testing kit.

Yet further, the proposed device 600, after the relevant DC and viscosity are measured for the edible oil being tested, may be configured to instruct the data processor 606 to compute in parallel the predicted values for the TPC and PTG of the edible oil. In other words, the predicted values for the TPC and PTG are calculated in parallel with simultaneous use of the aforementioned relevant mathematical models. It is also however to be understood that whether the predicted values of the TPC and PTG are computed in parallel or sequentially is not critical for the proper operation of the proposed device 600, so long as the computations for predicting the values of the TPC and PTG are performed sufficiently fast within a reasonable acceptable timeframe that would allow the proposed device 600 to be used as a "quick test kit".

It is to be appreciated that the respective coefficients in equations (1 ) to (3) as described above are provided simply for illustration purposes, and do not limit equations (1 ) to (3). Thus, other suitable combinations of coefficients that fit the described mathematical forms of equations (1 ) to (3) are also possible.

While the invention has been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered illustrative or exemplary, and not restrictive; the invention is not limited to the disclosed embodiments. Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practising the claimed invention.

Claims

1. A method of determining the quality of edible oil, the method comprises: determining a plurality of parameters of the edible oil, the plurality of parameters including dielectric constant and viscosity;

using the plurality of parameters as input variables of a mathematical model for predicting values of both the total polar compounds (TPC) and triglyceride polymeric content (PTG) of the edible oil;

comparing the predicted TPC and PTG values with corresponding predetermined thresholds; and

determining the quality of the edible oil based on the comparison.

2. The method of claim 1 , wherein the mathematical model includes a nonlinear mathematical model.

3. The method of claim 2, wherein the nonlinear mathematical model for predicting the value of the TPC is expressed as an equation: TPC = -12.353 - 0.0199DC² + 1.591 DC + 0.00519vis² - 0.0328vis,

where DC is the dielectric constant; and

vis is the viscosity.

4. The method of claim 3, wherein the equation is characterised with a root- mean-square deviation (RMSE) value of 2.01 , and with a R² value of 0.93, where R is the correlation coefficient.

5. The method of claim 2, wherein the nonlinear mathematical model for predicting the value of the PTG is expressed as an equation: PTG = -9.936 + 0.00608vis² - 0.0385vis - 0.00470DC² + 0.376DC,

where DC is the dielectric constant; and

vis is the viscosity.

6. The method of claim 5, wherein the equation is characterised with a root- mean-square deviation (RMSE) value of 1.50, and with a R² value of 0.92, where R is the correlation coefficient.

7. The method of claim 2, wherein determining the dielectric constant includes measuring the dielectric constant at a temperature of about 80°C.

8. The method of claim 2, wherein determining the viscosity includes measuring the viscosity at a temperature of about 40°C.

9. The method of claim 1 , wherein the plurality of parameters further include lightness, redness, blueness, absorbance and refractive index.

10. The method of claim 9, wherein determining the absorbance includes measuring the absorbance at a wavelength of 450 nm or 490 nm.

11. The method of claim 1 , wherein using the plurality of parameters as input variables of a mathematical model for predicting the values of both the TPC and PTG further includes computing the values of the TPC and PTG in parallel.

12. The method of claim 1 , further comprises generating an alert signal if the predicted TPC value or PTG value exceeds the corresponding predetermined threshold.

13. The method of claim 12, further comprises providing a visual indicator corresponding to the generated alert signal.

14. The method of claim 2, wherein the nonlinear mathematical model for predicting the value of the PTG is expressed as an equation: PTG = -73.208 +

0.00692vis² - 0.0413vis + 33.317RI² + 15.473RI,

where Rl is the refractive index; and

vis is the viscosity.

15. The method of claim 14, wherein the equation is characterised with a root-mean-square deviation (RMSE) value of 1.44, and with a R² value of 0.94, where R is the correlation coefficient.

16. A device for determining the quality of edible oil, comprising:

a plurality of probes adapted to determine a plurality of parameters of the edible oil, the plurality of parameters including dielectric constant and viscosity; and

a data processor adapted to:

use the plurality of parameters as input variables of a mathematical model for predicting values of both the total polar compounds (TPC) and triglyceride polymeric content (PTG) of the edible oil;

compare the predicted TPC and PTG values with corresponding predetermined thresholds; and

determine the quality of the edible oil based on the comparison.

17. The device of claim 16, wherein the data processor is further adapted to generate an alert signal if the predicted TPC value or PTG value exceeds the corresponding predetermined threshold.

18. The device of claim 17, further comprising:

a display screen for providing a visual indicator corresponding to the generated alert signal.

19. The device of claim 16, wherein the mathematical model includes a nonlinear mathematical model.

20. The device of claim 19, wherein the nonlinear mathematical model used by the data processor for predicting the value of the TPC is expressed as an equation: TPC = -12.353 - 0.0199DC² + 1.591DC + 0.00519vis² - 0.0328vis, where DC is the dielectric constant; and

vis is the viscosity.

21. The device of claim 19, wherein the nonlinear mathematical model used by the data processor for predicting the value of the PTG is expressed as an equation: PTG = -9.936 + 0.00608vis² - 0.0385vis - 0.00470DC² + 0.376DC, where DC is the dielectric constant; and

vis is the viscosity.

22. The device of claim 19, wherein the nonlinear mathematical model used by the data processor for predicting the value of the PTG is expressed as an equation: PTG = -73.208 + 0.00692vis² - 0.0413vis + 33.317RI² + 15.473RI, where Rl is the refractive index' and

vis is the viscosity.

23. The device of claim 19, wherein one of the plurality of probes is configured to measure the dielectric constant at a temperature of about 80°C.

24. The device of claim 19, wherein one of the plurality of probes is configured to measure the viscosity at a temperature of about 40°C.

25. The device of claim 16, wherein the plurality of probes includes at least two such probes.

26. The device of claim 25, wherein one of the at least two such probes is configured to measure both the dielectric constant and the temperature.

27. The device of claim 16, further comprising an input panel for reconfiguring values of the predetermined thresholds.