JPH11194124A - How to manage lubricants - Google Patents

Abstract

(57) [Summary] [Problem] Even for non-experts, it is possible to make accurate decisions in a short time on the specifications of lubricants, the timing of replacement with new bases, and the timing of replenishment of new bases and additives. To be able to do SOLUTION: Near-infrared spectra of a plurality of known lubricants including a base and various additives are measured, and properties, performance, etc. are predicted by multivariate analysis of the near-infrared spectra of the plurality of known lubricants. At the same time, the properties, performance, etc. of a plurality of known lubricants are measured, and model parameters of a calibration curve composed of the predicted values based on the plurality of known lubricant data subjected to multivariate analysis and the measured values are created. Then, the near-infrared spectrum of the unknown lubricant is measured, and properties and performance of the unknown lubricant are estimated from the measurement results and the model parameters of the calibration curve, and the unknown lubricant is managed based on the estimation result.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for controlling the properties and performance of a lubricant, and more particularly to a method for replacing a lubricant with a new base and a new base and additives. It relates to a method of accurately judging and managing etc.

[0002]

2. Description of the Related Art There are various types of lubricants depending on the application, such as lubricating oils such as engine oils and turbine oils, heat-treated oils and greases, and bases and additives differ depending on the types. Of course, the mechanism by which the lubricant deteriorates is also different. On the other hand, in using a lubricant, how to manage the specifications, when to replace deteriorated lubricant, or when to replenish new lubricant and additives are important issues.

[0003]

However, as described above, the types of lubricants are extremely large, and the additives are different for each type, and the deterioration mechanism is also different. I was For example, lubricants such as engine oil and turbine oil are deteriorated by use, and the oxidation stability is used as an index of the degree of deterioration. RBOT (JIS
K2514) is commonly used. Where R
In the BOT test, oxygen is pressurized and sealed in turbine oil, a catalyst is added, and the decrease time of the oxygen pressure is measured at 150 ° C., and the operation is complicated.

That is, a skilled person performs not only a physical property test but also a performance test according to the type of lubricant, analyzes these test results based on knowledge and experience cultivated over many years, and manages the specifications, exchange timing, new Judgment was made on the timing of base and additive replenishment. As described above, the conventional lubricant management method has problems that the number of test items is too large, the data analysis takes a long time, and the data analysis requires knowledge and experience.

[0005] The present invention has been made in view of the above problems, and even if a non-expert, the specification of lubricants, the timing of replacement with a new base, the timing of replenishment of the new base and additives, etc. It is another object of the present invention to provide a method for managing a lubricant that enables an accurate determination to be made in a short time.

[0006]

Means for Solving the Problems In order to achieve the above object, the invention according to claim 1 measures near-infrared spectra of a plurality of known lubricants containing a base and various additives, and measures these near-infrared spectra. Multivariate analysis of the near-infrared spectrum of known lubricants to predict properties, performance, etc., and actual measurement of properties, performance, etc. of multiple known lubricants. A model parameter of a calibration curve consisting of a predicted value based on the above and the actual measurement value is created, and then a near-infrared spectrum of the unknown lubricant is measured, and the properties of the unknown lubricant are determined from the measurement result and the model parameters of the calibration curve. This is a method of estimating performance and the like, and managing the unknown lubricant and the known lubricant based on the estimation result.
As a result, even non-experts can accurately estimate the properties and performance of unknown lubricants, and determine the specifications of lubricants, the timing of replacement with new bases, and the timing of replenishment of new bases and additives. Can be performed in a short time.

According to a second aspect of the present invention, the properties and performance of the lubricant are the degrees of deterioration of the lubricant, and RBOT values are obtained as the predicted value and the actually measured value to prepare a calibration curve of the RBOT value. The degree of deterioration of the unknown lubricant is estimated based on the calibration curve. Thereby, management according to the degree of deterioration of the lubricant can be performed.

According to a third aspect of the present invention, there is provided a calibration curve in which the properties and performance of the lubricant are the antioxidant amount of the lubricant, and the antioxidant amount is obtained as the predicted value and the actually measured value. The antioxidant amount of the unknown lubricant is estimated based on this calibration curve. This allows
Management according to the amount of antioxidant in the lubricant can be performed. The invention according to claim 4 uses principal component analysis and multiple regression analysis as the multivariate analysis.

In the present invention, the known lubricant is
During and after use, and unused lubricant,
The known lubricants are those whose properties and performance are known, and the unknown lubricants are those which use unused lubricants among the above-mentioned known lubricants, those which further use lubricants during and after use, and the above-mentioned known lubricants Lubricants whose physical properties, performance, etc. are unknown.

[0010]

Embodiments of the present invention will be described below. In this embodiment, a case where the present invention is applied to a method for testing the oxidation stability of a lubricant will be described. The method of managing the lubricant in this embodiment is generally performed according to the following procedure. That is, first, near-infrared spectra of a plurality of known lubricants are measured with a near-infrared spectrophotometer, and these are subjected to multivariate analysis, and properties and performance of the lubricants are actually measured.
The model parameters of the calibration curve of the RBOT value composed of the predicted value and the actually measured value are created. Thereafter, when managing the unknown lubricant, the RBOT value is predicted by measuring the near-infrared spectrum of the unknown lubricant, and the actual RBOT value of the unknown lubricant is calculated from the predicted value and the calibration curve created previously. Is estimated. Then, the lubricant is managed based on the estimated RBOT value.

Next, a method for managing the lubricant will be described in detail with reference to the drawings. FIG. 1 is a flowchart showing the procedure of the management method. First, the sample oil is measured by a near-infrared spectrophotometer. In this embodiment, the deteriorated lubricant oil is used as the sample oil (step S1). As the near-infrared spectrophotometer, any commonly used one can be used. For example, a monochromator using a quartz prism and a detector using a lead sulfide photoelectric tube is used. The near-infrared measurement wavelength range is 750 nm to 2500 nm
However, in the present embodiment, 1000 nm to 2100 nm
Is measured.

A near-infrared spectrum of a plurality of lubricants is measured by a near-infrared spectrophotometer (step S2). Here, a near-infrared spectrum from sample 1 to sample X is measured. This is schematically shown in FIG. 2, where y variables are obtained for each sample as the absorbance at each wavelength.

[0013]

The measured near-infrared spectrum is input to a computer and stored in a memory, and then sent from the memory to a calculation unit for performing multivariate analysis, and the calculation unit performs multivariate analysis (step S3). Specifically, analysis is performed by principal component analysis, and an axis is set in a direction in which data variance is largest, that is, a direction in which a difference in wavelength absorbance is largest, and is set as a first principal component. First principal component: Z ₁ = A ₁₁ N ₁ + A ₂₁ N ₂ +... + A _y1 N _y Substituting the values of the respective samples N _{1 to} N _y into the calculation formula of the first principal component, Z _{11 to} Z _The coefficient a _{11 is set} so that the variance of the value of _x1 is maximized.
To a _y1 . Sample 1 Z ₁₁ = a ₁₁ N ₁₁ + a ₂₁ N ₁₂ +... A _y1 N _1y Sample 2 Z ₂₁ = a ₁₁ N ₂₁ + a ₂₁ N ₂₂ +... A _y1 N _{2y. X Z x1 = a 11 N x1} + a 21 N x2 + ····· a y1 N xy

Here, the position of each sample on the principal component axis is referred to as a principal component Score, and is calculated by substituting N _{1 to} N _y of each sample x into the first principal component equation from which a _{11 to} a _y1 are obtained. This is the value of Z obtained. Since the value of a _{1 to} a _y is directly reflected on the principal component axis, a variable y (wavelength range) having a large value of a
Is greatly involved in the principal component axis. Values of coefficients obtained a ₁₁ ~a _y1 and Z ₁₁ to Z _x1 is stored in the memory.

Next, the value of Z of the second principal component is determined in the same manner as above so as to be in a direction perpendicular to the axis of the maximum direction of the variance and to maximize the variance. That is,
Seeking Z ₁₂ to Z coefficient a ₁₂ ~a _y2 such dispersion is the maximum value of _x2 in the following formula, and then determines the value of up to Z ₁₂ to Z _x2. Sample 1 Z ₁₂ = a ₁₂ N ₁₁ + a ₂₂ N ₁₂ +... _Ay2 N _1y Sample 2 Z ₂₂ = a ₁₂ N ₂₁ + a ₂₂ N ₂₂ +... _Ay2 N _2y. XZ _x2 = a ₁₂ N _x1 + a ₂₂ N _x2 +... A _y2 N _{xy In} this way, the coefficient and the Z, up to the n-th principal component, if necessary
Is obtained and stored in the memory.

Next, the RBOT value is predicted by PLS regression analysis (step S4). The computing unit of the computer reads the first to n-th main components from the memory,
Each principal component is analyzed by PLS regression analysis using the BOT value as an objective variable. _{Specifically, Y = B 0 + B 1} Z 1 + B 2 Z 2 + ··· B n Z n Y: value want to predict Z ₁ to Z _n: first to n principal component B ₀ .about.B _n: Amane Numerical regression coefficient n: number of optimal principal components for regression Z ₁ = a ₁ N ₁ + a ₂ N ₂ +... A _y N _y ... A _{1 to} a _y : coefficients N _{1 to} N _y : The RBOT value is predicted by the equation of absorbance at each wavelength y: number of variables.

Next, the RBOT value (R) of Sample 1 to Sample x
Is actually measured (step S5). Here, the actual measurement of the RBOT value is based on JISK2514 “Lubricant oxidation stability test method”.
Perform according to. Specifically, after a catalyst is put in a sample to a specified temperature and a specified pressure, a time until a certain pressure drop occurs is determined. The number of samples is required to be at least 20 or more, and preferably 50 to 100 in order to improve the reliability of the estimation formula. The more samples, the more accurate results can be obtained, but if it exceeds 100, there is a problem that overfitting is likely to occur. The variable y is usually 2 to 150 depending on the measurement wavelength.
It is about 0, and is 1100 in the present embodiment.

Next, a calibration curve of the RBOT value (R) is created (step S6). Specifically, a model parameter of a calibration curve as shown in FIG. 3 is created by plotting the measured value of RBOT on the X axis and the predicted value on the Y axis.

The RBOT value of the unknown sample is estimated (steps S7 to S9). The near-infrared spectroscopic measurement of the unknown sample is performed by the same method as described above (step 7), and the RBOT value is predicted (step 8). Then, based on the predicted RBOT value,
An actual RBOT value corresponding to the predicted RBOT value is obtained, and an actual RBOT value of the unknown sample is estimated (step S).
9).

Next, the physical properties, performance and the like of the unknown lubricant are estimated from the actual RBOT value, and the unknown lubricant is managed based on the estimation result (step S10).

In the above embodiment, the case where the oxidation stability of the lubricant is controlled has been described. However, it is possible to estimate the amount of the antioxidant, the total acid value, the viscosity, the water content, and the like by the same method. .

[0023]

[Examples] 1. Sample (a) Base mineral oil (viscosity 32 mm2 / sec, 40 ° C.) 98.4 wt% (b) Phenol antioxidant 1.0 wt% (c) Phosphorus antioxidant 0.5 wt% (d) Amine oxidation Inhibitor 0.1 wt% 2. 2. Near-infrared spectrometer: Infralyzer 500 manufactured by Bran Roubaix Measurement wavelength: 1000 to 2100 nm Multivariate analysis system: Unscra manufactured by CAMO
mbler The RBOT value of the sample was predicted using the device and system. 4. RBOT value measurement method: according to JIS K2514 A catalyst is added to the sample, oxygen is enclosed, and the oxygen pressure is 620 kPa.
And Copper was used as a catalyst, and the temperature was raised to 150 ° C. The oxygen pressure was increased from 620 kPa to about 1200 kPa. The time when the 175 kPa oxygen pressure decreased from the pressure at that time was measured starting from the temperature rise. Near-infrared absorption spectra of various oils were measured, and RBO
A calibration curve was created in comparison with the actual measured value of T value. Next, the unknown sample was measured, and the RBOT value was estimated using a calibration curve.
The results are shown below.

[0024]

Although the RBOT value is the most standard method for testing oxidation stability, the operation is complicated. The estimated value of the RBOT value by this method is in good agreement with the measured value,
With this method, the RBOT value can be easily estimated.

[0026]

According to the present invention, it is possible to easily estimate the physical properties, performance, etc. of an unknown lubricant from a calibration curve. An accurate judgment can be made in a short time about the timing of the replenishment of the new base and the additive, and the like. It is also possible to easily know when to replace the deteriorated lubricant or when to replenish new lubricants and additives.

[Brief description of the drawings]

FIG. 1 is a flowchart illustrating a method according to an embodiment of the present invention.

FIG. 2 shows a near-infrared spectrum from Sample 1 to Sample X measured by a near-infrared spectrophotometer.

FIG. 3 shows a calibration curve of RBOT value (R).

Claims (4)

[Claims] 1. The near-infrared spectra of a plurality of known lubricants including a base and various additives are measured, and the near-infrared spectra of the plurality of known lubricants are multivariately analyzed to determine properties, performance, etc. Predict and perform actual measurements of the properties, performance, etc. of a plurality of known lubricants, and create a model parameter of a calibration curve composed of the predicted values based on the plurality of known lubricant data subjected to multivariate analysis and the measured values, Next, the near-infrared spectrum of the unknown lubricant is measured, and properties and performance of the unknown lubricant are estimated from the measurement results and the model parameters of the calibration curve. A lubricant management method characterized by performing management. 2. The property and performance of the lubricant are the degree of deterioration of the lubricant, and an RBOT value is obtained as the predicted value and the actually measured value, and a calibration curve of the RBOT value is created.
2. The method according to claim 1, wherein the degree of deterioration of the unknown lubricant is estimated based on the calibration curve. 3. The properties and performance of the lubricant are the antioxidant amounts of the lubricant, and a calibration curve is prepared by calculating the antioxidant amount as the predicted value and the actual measurement value. 2. The method according to claim 1, wherein the antioxidant amount of the unknown lubricant is estimated based on the unknown. 4. The lubricant management method according to claim 1, wherein said multivariate analysis is a principal component analysis and a multiple regression analysis.