Chemistry and Technology of Fuels and Oils, Vol. 42, No. 1, 2006

METHODS OF ANALYSIS

FAST DETERMINATION OF HIGH WEAR OF FRICTION UNITS BASED ON PARTICLE SIZE AND CONCENTRATION Yu. A. Gur’yanov and N. I. Skinder

UDC 621.836:681.518.54(04)

Methods of detecting elevated wear of friction units in machines with the parameters of the wear particles in working lube oil are examined. The high sensitivity of the new fast method that allows analyzing wear particles in five minutes with simple and inexpensive equipment in field and laboratory conditions is demonstrated. The basic reason why different mechanisms reach the limiting state is wear of the working surfaces of friction units. This frequently occurs due to unsatisfactory technical maintenance of the machines based on a scheduled preventive maintenance system. In such a system, 34% of the machines are in a technically unacceptable state, 32% are acceptable, 21% are satisfactory, and only 13% are in good condition [1, 2]. In technically advanced countries, operative control of the condition is conducted to ensure the long lifetime of the machines, for example, internal combustion engines (ICE) for mobile technology whose run before the limiting state is 1-2.4 million miles (1.6-3.8 million km) [3, 4]. Technical diagnostic agents that determine the machine’s real maintenance and repair requirements in technical servicing are used for this purpose. Monitoring of working lube oils is one of the most effective methods of obtaining such information [2, 3], including revealing high wear of rubbing surfaces. The latter is determined with the parameters of wear particles in the working lube oil. The fundamental parameters are: the concentration of particles in the oil, their size and morphology, including shape, color, material, etc. High wear is usually judged with the size and concentration of wear particles while the morphology is analyzed in especially critical cases. The maximum particle size unambiguously characterizes the degree of severity of friction surface wear and the average value of the particle concentration characterizes the degree of elevated wear of the mechanism. In evaluating wear, it is also necessary to consider that perturbation of the standard friction regime in one of the interfaces usually does not result in any marked increase in the average value of the wear particle concentration in the initial stage, but can cause emergency shutdown of the mechanism. It is thus necessary to use the two particle parameters mentioned above for reliable diagnosis of high wear of working surfaces. This statement is ____________________________________________________________________________________________________ Chelyabinsk State Agroengineering University. State Scientific-Research Institute on Industrial Tractors. Translated from Khimiya i Tekhnologiya Topliv i Masel, No. 1, pp. 44 – 47, January – February, 2006. 60

0009-3092/05/4201–0060 © 2006 Springer Science+Business Media, Inc.

based on the functional dependence of the size and concentration of wear particles in working oil on the friction regime of bearing surfaces [5] (Fig. 1). As the friction regime becomes more severe, the wear particle size increases in logarithmic progression and the concentration increases in linear progression. As a consequence, the particle size is much more sensitive to a more severe friction regime than the concentration of particles in the oil as a diagnostic parameter. However, the combined use of these parameters in establishing a diagnosis allows significantly expanding the possibilities of recognizing the real condition of the mechanism. Analytical methods involving expensive equipment are primarily used for analyzing wear particle parameters: spectral units and ferrographs. Ferrographs are the best devices for analyzing wear particles since they determine both the concentration and the particle size by fractions. The spectral units allow determining the concentration of wear particles in oil only for sizes no greater than 10 mm, which essentially limits their sensitivity to changes in the friction regime. The basic drawback of the analytical instruments is their high cost. For example, a bare-bones ferrograph, the Industrial Tribology System from SPECTRO INC. (USA), costs $50,000 according to information from the company. There are also methods based on passing working oil through a filter consisting of several layers with pores of different sizes [3-5]. These methods allow determining the fractional particle size and concentration with a known pore size based on the mass of the ferromagnetic particles precipitated in each layer of the filter [6-8]. The need for company-manufactured filters is their basic drawback. There are also simple magnetometric devices for determining the concentration of ferromagnetic wear particles in oil. However, diagnosis of mechanisms based on the concentration of particles usually does not ensure unambiguous recognition of high wear of friction units in the early stage. A new fast method which has the abilities of analytical methods is proposed in [9]. It uses a simple, inexpensive, and portable instrument that determines the wear particle size and concentration in 5 min in field and stationary conditions without expendable materials.

Concentration, rel. units

IV

III

10 1

2

3 II

5 1 0.1 1

I 10 100 1000 10 000

Fraction of particles, µm

Fig. 1. Change in concentration and size of wear particles on working surfaces: I, II, III, IV) in standard, high, prebreakdown, and breakdown wear; 1) detected by the spectral method of diagnosis; 2) same but with special adaptation; 3) not detected by the spectral method of diagnosis. 61

We demonstrated the effectiveness of this fast method for diagnosis of mechanisms operating in a liquid grease system based on the parameters of metal wear particles in the lube oil for detecting elevated wear of friction units in the early stages. A hypothesis was first advanced concerning the possibility of reliable estimation of the wear particle parameters with the fast method based on synthesis of two existing methods: sedimentometric and magnetometric [10]. Sedimentometric methods are based on Stokes’ dependence, which allows calculating particle radius R with the particle sedimentation rate θ in the liquid:

θ = 2 g (ρp − ρl )R 2 / 9η where g is the free fall acceleration; ρ p, ρ l are the density of the particle material and liquid; η is the dynamic viscosity of the liquid. After experimentally determining (i.e., during diagnosis) the particle sedimentation rate with magnetometric devices, the particle size can be calculated with the above equation for ρ π, ρ l , and η = const. The quadratic dependence between the particle sedimentation rate and particle radius gives the method high basic sensitivity. A difference in the densities of the wear particle material, iron, for example, and the lube oil by 8.7 times additionally increases the sensitivity of the method by several times. The functional diagram of the device for analyzing wear particle parameters and the possible versions of assembly of the sensitive element (electromagnetic sensor) in measurements is shown in Fig. 2. Figure 3 shows the qualitative change in the sensor current during sedimentation of wear particles in a sample of working oil in standard operation of the friction units in the diagnosed mechanisms and in increased wear. A sample (30-50 ml) of oil from the crankcase is collected for diagnosis with a sampler 40-70 mm deep immediately after stopping the mechanism which was previously operating in the usual regime and placed in a plastic or glass container. If the analysis of the wear particle parameters is performed immediately after collecting the sample, the sensor is immersed in the tested oil, for example, according to version III (see Fig. 3), and the value

Version II Version I 1

2

Sensor Sensor Power unit

Signal processing unit

Recording unit

Sensor Äŕň÷čę Version III

1

Fig. 2. Variants of the diagram of the measuring device for determining the sedimentation rate of magnetic wear particles: 1) oil sample; 2) clean oil with analyzed sample added. 62

I, mA

of the current I n which does not change for several seconds is taken. Then the change in the readings is observed and the second current value I c is taken after 5 min. The first value I n characterizes the average integral concentration of metal particles in the tested oil and corresponds to the initial horizontal segment of the curves. The change in the current ∆ i after time t referring to t, i.e., parameter tan α, is an indicator of the precipitation rate of wear particles from the largest fraction and consequently their size. The latter also allows unambiguously evaluating the real friction regime and wear of the working surfaces of parts of the diagnosed mechanism. The descending curves in Fig. 3 correspond to this process. If the sensor is installed according to version I or II (see Fig. 2), the dependence is the mirror image in the graph relative to the abscissa. The initial horizontal segment is absent, and the concentration of wear particles is estimated with the maximum value of the current corresponding to the horizontal segment at the end of the sedimentation curve. If the analysis is performed as some time elapses, the sample should first be heated to 50-60°C and carefully mixed by shaking. The maximum wear particle size in standard operation of mobile technology aggregates for general applications is 10-20 mm, 100-250 mm in the initial state of increased wear, and in breakdown, from one to several

∆2

α

2

∆1

1 t, minutes

t = 5 minutes

Fig. 3. Change in current I in time at a sensor position according to version III: 1) in standard operation of friction units; 2) with large wear particles in the working oil; ∆ i = I n – I c is the change in the current during the measurement with t = 5 min; tanα = ∆ i /t is a parameter characterizing the wear particle size in the largest fraction.

−4

Concentration, 10 %

160 2

120 3

80

4

1

40 0

20

40 60 Running time

80

100

Fig. 4. Concentration of iron particles vs. oil running time in test stages: 1) 1-10; 2) 10-20; 3) 20-30; 4) 30-40. 63

mm. As a consequence, in the initial stage of evolution of high wear and the appearance of 100 mm particles in the working oil, the diagnostic parameter of particles, even the lightest duralumin particles, increases by one order of magnitude at the minimum. The theoretical and experimental studies and practical experience in diagnosis of ICE and transmitting aggregates showed that as the wear regime progressed from standard to breakdown, the particle sedimentation rate increased by three to four orders of magnitude and the concentration increased by one order of magnitude. This confirms the hypothesis concerning the high sensitivity of the method in detecting elevated wear. To increase the reliability of diagnosis, it is useful to measure the temperature of the oil sample in analyzing the particles and to take it into consideration in terms of the viscosity of the oil in making the diagnosis. A s a n e x a m p l e , w e r e p o r t t h e r e s u l t s o f d i a g n o s i n g a V- 8 4 d i e s e l e n g i n e d u r i n g o p e r a t i n g tests over 4500 h on a Schenck bench (Germany) at the Scientific-Research Institute on Industrial Tractors (SRI IT). Samples of working oil from the ICE crankcase were collected after 10 motor hours and the oil was replaced after 100 motor hours. The quality of the working oil was monitored in the certified laboratory at SRI IT based on the physicochemical properties and the friction and wear regime of the working surfaces of ICE parts was monitored by the spectral method based on the concentration of iron and also by the fast method examined. The experimental dependences of the concentration of iron in working oil as a function of the running time of the oil determined by the spectral method are shown in Fig. 4. The periodicity of monitoring the concentration of iron particles in the oil was 10 motor hours in stages 1 to 10 and 50 motor hours in stages 10 to 40. The character of the curves of these dependences is the same. Wear particles gradually accumulated after it was changed and the intensity decreased after 40-50 h of operation. The difference between the curves is within the limits of the error of the method. In standard operation of the ICE, the character of the change in the concentration of ferro-, para-, and diamagnetic metal particles in the oil with the running time, measured with an electromagnetic sensor, is the same (Fig. 5) as with the spectral method. However, ad hoc perturbation of the standard friction regime in stages 30 and 31 in one of the “duralumin–steel” bearings (there are more than 50 in ICE) led to high wear of this bearing, the appearance of large duralumin particles in the oil, and an increase in their average concentration.

Stage 35

8

4

0

Concentration increase in 5.9 times

Current, mA

Stage 32 Stage 31

2

Stage 24 1

20

40 60 Running time

80

100

Fig. 5. Current characterizing the concentration of duralumin wear particles as a function of ICE running time: 1) in standard regime; 2) in breakdown regime; the dashed segment on curve 2 corresponds to breakdown failure, after which the ICE was serviced, the centrifuge was cleaned, and 50 liters of oil were added. 64

17

Current, mA

3 13

2 9

1 5

0

3 6 Time, mintes

Fig. 6. Duralumin wear particle precipitation curves: 1) y = 8.6453e -0.0431x in stage 24; 2) y = 17.318x -0.4547 in stage 31; 3) y = 17.249e -0.0665x in stage 35. In stages 31-33, the increased wear regime gradually turned into the breakdown wear regime. As a result, after 10 h of running fresh oil, the concentration of metal particles in it at the end of stage 31 was 6 times higher in comparison to the standard value. The high concentration of particles in stage 35, i.e., after maintenance, cleaning of the oil centrifuge rotor, and adding fresh oil to the ICE’s crankcase, was due to the important amount of duralumin particles in the oil that remained on the surfaces of the parts of the crank mechanism and on the inner walls of the crankcase and drip pan. The initial fragments of the experimental sedimentation curves for stages 24, 31, and 35 are shown in Fig. 6. for comparatively assessing the sensitivity of the method to the maximum wear particle size. The analysis of the graphs suggests that particles larger than in stages 35 and 24 appeared in the working oil in stage 31. In stage 31, the particle size index was tan α = 3 at t = 1 min and 1.2 at t = 5 min; in stage 35, 0.7 at t = 5 min; in stage 24, 0.3 at t = 5 min, and 0.3 at t = 3 min. In stage 35, the concentration of wear particles was higher than in stage 31, for the previously mentioned reason. However, the particle size was one order of magnitude smaller due to standard operation of all ICE bearings after maintenance. When high wear is detected, it is useful to reduce the measuring time to 30-60 sec. This allows more accurately determining the maximum particle size as a result of narrowing the fraction used to determine the average size. In stage 31, as a result of reducing t from 5 to 1 min, the particle size index tan a increased by 2.5 times. The high efficacy of recognizing increased wear by a simple fast method implemented on an equally simple instrument was thus confirmed by the results of diagnosing ICE with the size and concentration of wear particles in working oil. When the spectral method was used, increased wear was not detected. This was due to monitoring with iron particles whose average concentration in the oil varied within the limits of the error of this method, not determining the concentration of duralumin particles, and the insensitivity of the method to large particles (see Fig. 1). The breakdown situation in ICE was not reflected in the physicochemical properties of the oil. 65

REFERENCES 1. K. Yu. Skibnevskii, G. Ya. Tel’nova, I. A. Yansikene, et al., Tekhnika v Sel’sk. Khozyaistve, No. 4, 45-48 (1979). 2. A. V. Lavrov and S. M. Prytkov, Gorn. Promyshl., No. 5, 38-39 (2002). 3. J. A. Geehan and P. R. Ryason, Million Mile Bearings: Lessons from Diesel Engine Bearing Failure Analysis, www.cospic.ru, p. 11. 4. B. Gorbunov, Gorn. Promyshl., No. 5, 20-21 (2001). 5. S. Bubenicek, Bull. Techn. Bureau Veritas, 75-78 (March/April 1987). 6. USSR Inventor’s Certificate No. 431451. 7. B. J. Roylance, M. H. Jones, and A. L. Price, in: Materials Performance and Conservation Conf. Swansea (1978), pp. 119-125. 8. E. M. Markov, L. S. Pinchuk, V. A. Gol’dade, et al., Trenie Iznos, 16, No. 3, 518 (1995). 9. Yu. A. Gur’yanov, Tekhnika v Sel’sk. Khozyaistve, No. 3, 30-33 (2000). 10. RF Patent No. 2082150.

66