資源目錄里展示的全都有，所見即所得。下載后全都有,請放心下載。原稿可自行編輯修改=【QQ：401339828 或11970985 有疑問可加】 Prototype fretting-wear testing machine and some experimental results Etsuo Marui *, Hiraki Endo, Norihiko Hasegawa, Hidetoshi Mizuno Department of Mechanical and Systems Engineering. Facility of Engineering. Gifll Unh-ersity. 1-1 Yanagido. Gifll-shi 501-11. Japan Received 8 July 1997; accepted 8 October 1997 Abstract A new prototype fretting-wear testing machine is developed, and its performance is evaluated. In this testing machine, a spherical specimen is oscillated by a lead-zirconate-titanate semiconductor (PZT) actuator, driven by means of computer control. Thus, an arbitrary input waveform, such as sinusoidal wave, triangular wave and so on, can be easily operated to a spherical specimen. This is the special feature of our testing machine. As an example, a fretting experiment is carried out, using a steel flat specimen and a bearing steel sphere specimen. A sinusoidal input wave is added by a PZT actuator. It is clarified that there is a boundary fretting amplitude, beyond which fretting wear mass or volume increases abruptly. ? 1998 Elsevier Science S.A. Ke.nmrds: Fretting wear; Fretting-wear testing machine; Sphere-an-flat type; PZT actuator 1. Introduction Fretting occurs wherever small-amplitude reciprocating sliding between contacting surfaces is sustained for a large number of cycles. The sliding amplitude or fretting amplitude is smaller than the contact area size and wear debris is likely to build up within the contact zone, so the surface conditions are apt to affect more remarkably the fretting wear phenomena than the usual wear phenomena .The main focus of previous research on fretting wear has been to basically understand its characteristics by clarifying the slip within the contact zone and the stress distribution there. Many experiments have been carried out to clarify the fretting wear characteristics of various engineering materials for practical use. The tested materials include mild steel , eutectoid steel , bearing steel , stainless steel, Cr-Mo steel , aluminum alloy , titanium and nickel alloys , and reinforced composite materials. The effects of surface oxidation ,lubrication and surface integrity by plasma or laser treatment have also been investigated. The following trend has been revealed by surveying the specimen configuration and the way in which specimens are oscillated. Specimen configurations used to date are: sphereon- flat surface type , cylinder-on-flat surface type and flat surface on flat surface type. Spheres or cylinders are oscillated in somecases, and a flat surface in others, on a sphere-on-flat surface type or cylinder-on-flat surface type testing machine. Means to oscillate specimens are: mechanical devices such as eccentriccam mechanisms ,hydraulic actuator mechanisms and electro-magnetic devices. A testing machine to observe the effect of environmental parameters such as temperature has also been constructed. Basing on the previous research, we designed and constructed a new prototype testing machine for fretting wear in this research program. Our testing device is inexpensive, and the testing conditions can be easily altered. Some experimental results of steel sphere-an-steel flat surface are reported here. 2. Prototype fretting-wear testing machine 2. J. Construction of machine Fretting wear is apt to occur when the reciprocating sliding amplitude between two surfaces ranges from several to tenodd J.lm [I]. Fretting wear is influenced by reciprocating sliding amplitUde, normal force, frequency of reciprocating movement, fretting cycle number, environment and physical property of materials. Thus, a fretting-wear testing machine, in which these influencing parameters can be altered easily and over a wide range, is favorable in practical use. In our prototype testing machine, a lead-zirconate-titanatesemiconductor (PZT) actuator, which can generate a smallre ciprocating displacement by a piezoelectric phenomenon[22], is used as a driving element for reciprocating sliding specimens. The PZT actuator is very advantageous becausevarious waveforms of reciprocating sliding movements canbe easily realized by computer-controlled voltage to the PZT actuator. Some of the electrical and mechanical properties of PZT actuator are given in Table 1. A combination of conventional spherical and flat specimens (sphere-on-flat type) are adopted in our machine. An outline of the fretting-wear testing machine used is given in Fig. I. The machine must be as rigid as possible to minimize the absorption of the small reciprocating sliding, which is induced by the PZT actuator, by an elastic deformation of each machine component. To realize this require- ment, the spherical specimen is directly attached to the drive section of the PZT actuator. For this reason, a spherical specimen 20 mm in diameter made of bearing steel is used. However, a spherical specimen of a different size can be readily used in this testing machine, without any modification of the testing device. Table 1 Some properties of PZT actuator Properties of PZT Relative dielectric constant 1800 Loss tangent 0.02 Young's modulus 44GPa Curie point 230°C Maker Nippondenso Size Diam'eter 20mm Length 120mm Specification Maximum displacement 100um Maximum driving force 2kN Maximum operating frequency 100Hz Allowable temperature limit 0-50°C Control voltage 0-5 V Fig. I. Prototype fretting-wear testing machine. A spherical specimen was arranged so as to be sandwiched between two identical flat specimens. By introducing this structure, a moment established owing to the friction force I acting on the spherical specimen can be eliminated, and the I : spherical specimen can be very rigidly held in place. The. driving clement and the sphere specimen are arranged on a ; vertical straight line, so the testing machine takes up little space. Moving elements arc suspended by two sets of parallel leaf springs as seen in Fig. 2, so that the moving direction of the PZT actuator parallels the contacting surface between specimens. Normal load is applied by the two dead weights shown in Fig. I, and the load can be thus kept constant during the experiment. The magnitude of the displacement generated by the PZT actuator is controlled through a personal computer. Control voltage of 0-5 V, generated by a personal computer, is magnified 120-fold through an amplifier. Then, the driving voltage of the PZT actuator is between 0-600 V. Computer control of the fretting-wear testing machine enables the spherical specimen to be driven by an arbitrary waveform. Two flat specimens 35 mm-square are attached to each stage, which can move horizontally owing to linear guides and rails. Wires are stretched between stages via pulleys, and the tensile force of dead weights is exerted on the stages. A spherical specimen is suspended between two identical flat specimens, and a constant normal force is added between specimens. Two sensors of leaf spring construction are attached on stages to measure the reciprocating sliding movement of specimens. Their size and configuration are given in Fig. 3.Each sensor consists of a leaf spring and a base. The base of the sensor is fixed to the stage. The tip of the leaf spring is arranged to be right above the spherical specimen. The deformation of the PZT actuator or the relative reciprocal sliding bet\veen spherical and flat specimen is measured by strain gauges pasted on both sides of the leaf spring. To minimize noise, all lead wires are sheathed. The calibration test result of these two sensors is shown in Fig. 4. The relation between output voltage from strain gauges and relative reciprocal sliding is shown here. Good linearity is recognized between them. Output noise is about 0.01 mY, which corresponds to the reciprocal sliding of 0.5 fLm. Accordingly, the relative reciprocal sliding between specimens can be measured precisely by these sensors. Fig. 2. Driving mechanism of specimens Fig. 3.Leaf Spring sensor to measure relative reciprocal sliding 2.2. Performance of prototype fretting-wear testing machine Performance of our prototype fretting-wear testing machine is checked through various calibration tests. Firstly, Fig. 5 gives a change with the passage of time of fretting amplitude. In this experiment, the input voltage waveform to the PZT actuator is sinusoidal, with an amplitUde and frequency of360 V and 10 Hz, respectively. In the figure, the result for the case of a bearing steel spherical specimen and a medium carbon steel flat specimen is shown at four levels of normal force. This combination of materials is used in the under-mentioned fretting-wear experiment. The fretting amplitude decreases rapidly, during the initial 2000 fretting cycles. This decrement of the fretting amplitude is more remarkable as the normal force becomes larger. Surfaces of both specimens are fully degreased by washing in methyl acetate before the fretting-wear test. Rapid decreasing of the fretting amplitude in the initial stage may be induced by the change of surface oxide film or the breakdown of slightly remained lubricant film. After that, the variation in fretting amplitude is recorded till the fretting cycle number of 105 .It is confirmed that the fretting amplitude is almost constant in relation to the magnitude of the normal force. Fig. 4. Calibration result for Fig. 5. Variation of fretting amplitude owing to fretting cycle leaf spring sensor. number at input voltage amplitude 360 V and frequency 10 Hz. The relation between the mean fretting amplitUde and the normal force from 3000 to 104 fretting cycles is shown in Fig. 6. The fretting amplitude decreases almost linearly with the increase of the normal force, and the relation between them is expressed by Eq. (I). A=-0.24Fn + A0 （1） where, A is the fretting amplitude (fLm), andAo is the fretting amplitude (fLm) when the normal force is zero. Fn indicates the normal force (N). The fretting amplitude decreases when the normal force increases, because\the frictional force acting between specimens grows as the normal force increases. Therefore, the frictional force can be estimated by the size of the fretting amplitude. Next, the relation between the mean fretting amplitude and the input voltage amplitUde is given in Fig. 7, for constant input voltage frequency (to Hz). In the figure, the result is given for two levels of normal force, 0 Nand 66 N. The mean fretting amplitude changes linearly with input voltage amplitude, and the lines indicating the relation between them are parallel to each other. It is ascertained from this that the decrease in mean fretting amplitude owing to the frictional force effect is not affected by the input voltage amplitude. From Fig. 7, the mean fretting amplitudeAo when the normal force is 0 N, is represented by Eq. (2). A0 = 0.13Va - 15 （2） where, Va is the input voltage amplitude (V) to the PZT actuator. From Eqs. (1) and (2), the fretting amplitude can be written as follows: A = 0.13Va - 0.24Fn - 15 （3） Fig. 6. Influence of normal force on fretting Fig. 7. Influence of input voltage magnitude to PZT amplitude at input voltage actuator on fretting amplitude at amplitude 360 V and frequency 10 Hz. input voltage frequency 10 HZ Lastly, the effect of the input voltage frequency is investigated. Fig. 8 shows the result for the case of the normal force 66 N and the input voltage amplitude to the PZTactuator 600 V. The mean fretting amplitude is almost constant tiII the frequency of20 Hz. Above this frequency, the fretting amplitude decreases with the input voltage frequency. From this result, it is ascertained that the limiting input voltage frequency of this fretting-wear testing machine is about 20 Hz. 2.3. Estimation of frictional force This fretting-wear testing machine is not equipped with asensor to measure frictional force. The magnitude of the frictional force can be approximately estimated as follows. As mentioned in Fig. 6, the frictional force is calculated using the difference between the fretting amplitude Ao when Fig. 8. Influence of input voltage frequency to PZT actuator on fretting amplitude at normal force 66 N and input voltage amplitude 600 V. the normal force is zero and the fretting amplitude A when the normal force is applied. The action of the frictional force is shown in Fig. 9, when the sphere specimen is sliding in the direction M as seen in the figure. The downward frictional force 2Q acts on the spherical specimen and the upward,. frictional force Q acts on each flat specimen and stage. The rigidity coefficient of the driving element in this direction is set at Rd and that of the stage at Rs. The variation or the decrease ilX of the fretting amplitude due to frictional force' is written as follows: 2Q/Rd + Q/Rs =ΔX （4） Estimation of the rigidity coefficients is carried out as follows: The part of the spherical specimen attaching to the driving element is pulled by a spring balance, and the displacement of the part attaching the spherical specimen is read by a dial gauge as seen in Fig. 10. The rigidity coefficient of the driving element Rd is then calculated by the pulling force and the displacement there. The rigidity coefficient of the stage Rs is obtained by the same method. As a result, the rigidity coefficient of the driving element may be stated Rd= 16.54 N/v.m, and that of the stage by Rs=8.7 N/v.m. Then from Eq. (4), the frictional force Q is obtained as follows: Q = 4.24 ΔX （ 5） The frictional coefficient is also estimated by dividing this frictional force by the normal force Fn. The frictional coefficient variation obtained by this process is given in Fig. II. In this estimation, the input voltage frequency is 10Hz, its amplitude is 480 V and the applied normal force is 66 N. The stable frictional coefficient is estimated as seen in the figure, except the initial stage of the fretting cycle number. Fig. 9. Frictional force acting on specimens. Fig. 10. Evaluation of driving element rigidity. Fig. II. Example of frictional coefficient evaluation at normal force 66 N. input voltage amplitude 480 V and frequency 10Hz. 'Fig. 12. Measurement of fretting wear scar at normal force 66 N. fretting frequency 20 Hz, fretting amplitude 35 ILm and fretting cycle number 105. 2.4. Estimatioll of wear volume Fig. 12 shows the fretting-wear scar in the case of the bearing steel sphere specimen and the medium carbon steel flat specimen. This is done by combining the surface roughness profiles measured by a surface roughness tester. Here, the applied normal force is 66 N, the input voltage frequency is 20 Hz, the fretting amplitude 35 fJ.m and the fretting cycle number 105 Fig. 12a shows the wear scar of the flat specimen and Fig. 12b shows the wear scar of the spherical specimen. The wear vestiges on the flat specimen have a conical shape. Its radius is about 600 fJ.m and the depth is about 14 fJ.m. There is a small plateau at the center of the worn portion, whose radius and depth account for about one-third of the wear scar. Around this area, there is an annular hump along the edge about 2.5 fJ.m high and around 100 fJ.m wide. On the other hand, the wear scar on the sphere specimen is a little smaller than the scar on the flat specimen. Its radius is about 570 fJ.m. There is a plateau at its center, similar to the flat specimen. The surface of the wear scar is almost smooth, except for the central plateau.The above features are similar to those of the fretting-wear scar already reported by many resf;!archers. Thus, the wear phenomena, which are observ,ed in o,ur fretting-wear testing machine, are considered to be induced by fretting wear. The fretting-wear volume is obtained [10] from the mean wear scar radius r and the scar depth d, which are easily measured by a surface roughness tester. Assuming the shape may be regarded as a sphere crown, the wear volume Vr is written as follows: Vr= d/6（3r^3+d^3) （6） It is extremely difficult to measure the wear scar depth d on the sphere specimen from a surface roughness curve. So, the radius of curvature of the wear scar surface is approximately regarded as the radius of the spherical specimen and the wear scar depth is calculated by Eq. (7). The approximate wear volume of the sphe~ical specimen is obtained by substituting Eq. (7) into Eq. (6). （7） where, Rand r represent radii of the sphere specimen and the mean wear scar radius, respectively. 3. Some experimental data on pair of stcel specimcns A fretting-wear test is carried out using the bearing steel sphere specimen and the flat specimen of medium carbon steel for machine construction, and the validity of our testing machine is ascertained. The chemical composition and Vick-ers hardness ofthe specimens are listed in Table 2. The radius of the spherical specimen is 20 mm. The flat specimen is a square plate 10 mm thick and 40 mm long. The fretting-wear test is carried out while varying some parameters affecting the fretting-wear characteristics. 3.1. Influence offretting cycle number The influence of the fretting cycle number is examined, while varying the fretting cycle number between 103 and 106 for the constant normal force 66 N, the fretting frequency 20 Hz and the fretting amplitude 35 p.m. A sample experimental result is shown in Fig. 13. The fretting-wear rate of the sphere and the flat specimens, which is converted from the wear volume obtained by the abovementioned process, is shown here. The fretting-wear rate is defined as the wear volume corresponding to the unit sliding distance and unit normal force. The concept of the wear raten is important to compare the wear quantitatively under different fretting-wear conditions. The wear rates for the spherical and flat specimens are similar in magnitude. As is clear in the figure, the wear rate becomes small with the increase of the fretting cycle number. There is a linear relationship between the wear rate and the fretting cycle number on log-log paper. This relationship is written as follows: （8） Fig. 13. Influence of fretting cycle number on fretting wear rate at normal force 66 N, fretting frequency 20 Hz and fretting amplitude 35 f.l.m. 4. Observation of wear scar by SEM It is clarified in the previous discussion that there is a boundary amplitude in the flat specimen. Then, by observing the wear scar surface by an electron microscope (SEM), the mechanism of the fretting wear is considered basing the stress on the relation with th~ fretting amplitude. 5. Conclusion A prototype fretting-wear testing machine, in which a sphere specimen is oscillated by the PZT actuator, is constructed and its perfonnance is evaluated in this paper. The most advantageous feature of this testing machine, in which a specimen's reciprocal sliding displacement is driven by the PZT actuator controlled by a personal computer, is that the specimen's reciprocal sliding displacement of sinusoidal, triangular wave fonn or others can easily be realized. The relation between the experimental conditions and the fretting amplitude is clarified in detail, considering the PZTactuat