Abstract
Reliability refers to the ability of a part,device or system to conduct an intended function in a given condition for a certain period of time.Amechanicalsystemorstructuresuchasamachinetoolexercisesthecapacity of the entire system with regard to the various constituent parts that are connected to each other; as such, the reliability of the parts constituting the system determines the reliability of the entire system. A tool post is a device designed to efficiently provide the tools necessary for the processing of a turning machine:the parts used in a hard turning machine which requires higher stiffness must provide greater reliability. For the purposes of this study, the reliability of a tool post, which has the highest failure rate of a turning machine system, was assessed. In order to conduct a reliability assessment of a given tool post,reliability prediction using a failure rate database,weak point analysis, the manufacture of a reliability tester and the calculation of reliability testing and quantitative reliability criteria were also carried out. By so doing, the failure rate, the MTBF (Mean time between failures) and other factors could be calculated. Furthermore,the results can also be applied to otherpartsoftheturningmachineortoareliabilityassessmentofasubsystem by using the suggested assessment method.
Keywords: Reliability assessment; Reliability prediction; Failure rate database; Tool post; Mean time between failures; Failure rate
1. Introduction
A production method to which the concept of reliability is applied has recently been used, rather than simple design and production focusing on the functions in all industrial fields (Saleh, 2006). Reliability refers to the ability of a part, device or system to conduct an intended function in a given condition for a certain period of time. Products that are produced according to such a method meet the customers' requirements in terms of both quality and function. In particular, a mechanical system or structures like a machine tool which exercises the capacity of an entire system in which the many constituent parts are connected to each other; as such, the reliability of the parts constituting the system determines the reliability of the entire system. Therefore the reliability of each part is very important (Lee, 2006). A tool post is a device that efficiently and automatically provides the tools necessary for the processing of a turning machine, and the precision of such a device is the core unit that ultimately determines the precision of a processed product. According to the relevant analyses, a tool post is known to have the highest failure rate among the subsystems that constitute a turning machine system
(RAC, 1991).In particular,a tool post used in relation to processing in a processing system - which requires high stiffness, such as that provided by a hard turning machine-requires higher reliability (Kim, 2005). In general, the reliability assessment of the electronic parts is conducted on the assumption that the failure rate (according to the bath-tub failure rate curve which is generally used) remains the same during the useful life of the electronic parts (Lee, 2001; Lee, 2006). However, although the failure rate of the mechanical parts tends to increase, it is an essential to obtain as much information on the reliability of mechanical parts as possible, all the more so because we don't currently have much information on the failure rate of mechanical parts (Wang, 1999; Lee, 2003). In this study, with regard to the reliability assessment of the tool post used in a hard turning machine, the quantitative calculation of reliability and the calculation of the reliability information of the mechanical parts were conducted by forecasting their reliability and analyzing their weak points using the failure rate of the mechanical parts;manufacturing a reliability tester for the reliability testing of a tool post; carrying out a reliability test for the measurement of such functions as stiffness, repetition and angular resolution; and calculating the quantitative reliability criteria, and so forth.
2. Reliability prediction
Reliability prediction refers to the efforts made to enhance the competitive power of a product in the market and to prevent losses caused by unexpected accidents, mainly by checking the reliability of the product's design according to its development state or by forecasting the reliability of a prototype, thereby enhancing its reliability before production starts (Moasoft Inc., 2002). The reliability prediction methods include FMEA (Failure Mode and Effect Analysis), FTA (Fault Tree Analysis), Worst Case Analysis, performance assessment and the field data method (Customer Service Data), and the failure rate database method, and so on. To conduct reliability prediction effectively,data (information on the failure rate) on the failures of each part is desirable. Unlike electronic parts, there is no clear definition of the failure mode and known reliability data for mechanical parts. Therefore, in this study,we conducted a reliability prediction using the NPRD95 (Nonelectric Part Reliability Data 95), a database containing information on the failure rate of mechanical parts(Lee, 2003). The NPRD95 database, which holds collected and edited data accumulated from 1974 to 1994, is the only source of information on the failure rate of mechanical parts. These failure rates follow an exponential distribution (RAe, 1995). In order to search for information on reliability,modeling of the system is required first of all.The basic data for modeling are the parts list,bills of materials and drawing, and so on. Once modeling has been completed, the reliability information should be entered using the failure rate database. For reliability information, the user chooses the failure rate under the usage environment in the part selection set to part, part sub-type. Figure 1 shows an example of a search of the Connector Pin for the failure rate by using the NPRD95. A tool post, as shown in Fig. 2, is composed of a turret head to which the tools are mounted, a main shaft that supports the turret, a clamping part to fix the rotation of the tool, gears (drive shaft) to transmit power for the rotation of the tool, and electrical sensor parts such as a proximity switch. The turret head has a Fig. 3.Analysis of the sub-assembly and main parts of the tool post.
Time Fig. 4.. Reliability change of the tool post and sub-assembly over time.
drive-gear, whose failure rate is 42.411000 failures/ million hours, while the lowest failure rate is found in the electricity parts, whose failure rate is 19.687800 failures/million hours. The failure rate being in inverse proportion to the MTBF in an exponential distribution, the result means that the mean time between failures of the drive gear is shortest and the failure rate of the electricity parts is longest. Figure 3 illustrates the failure rate of the sub-assembly of the tool post and the weak points and failure rate of the main parts. The percentage of each blank indicates the failure rate of each sub-assembly when it is assumed that the failure rate of the tool post is 100%. The Timing Belt, Pulley and Radial Bearing of the Drive Gear, which has the highest failure rate, have a high failure rate and therefore can be expected to be weak parts. As they also have a high failure rate, the Proximity Switch, Quad-Ring (X-Seal), 3 piece type
holder into which 12 tools can be inserted and installed. Because most mechanical products are composed of components that are linked to each other by rings, bolts and nuts, we classified the composition of the tool post to a single level for reliability prediction. The reliability information should be searched for according to the standards of the specifications, materials, and usage environment of the constituent parts. For the material-related specifications, we referred to the specifications manual of KS 0430I gray cast iron products and KS 03709 nickel chrome molybdenum steel materials, while for parts-related specifications, we referred to the KS specifications and in-house specifications standard. Because the desired usage environment and specifications of the constituent mechanical parts are not always available, we selected the most similar parts (usage environment, materials and specifications) in consultation with the designer. A reliability block diagram is a method for calculating failure rate-related reliability by expressing the flows of energy,matter and information shown by the system (Wang, 2004). In this study, where the tool rotation of the tool post is regarded as the main function, the main parts were put together by series connection. For the prediction results, the MTBF of the tool post was estimated at 8,590 hours and the failure rate at 116.408200 failures/million hours. The reliability prediction conditions involved an operation temperature of 30°C in a GB (Ground Begin) and GC (Ground Controlled) environment.
With regard to the sub-assembly, the highest failure rate is found in drive-gear, whose failure rate is 42.411000 failures/ million hours, while the lowest failure rate is found in the electricity parts, whose failure rate is 19.687800 failures/million hours. The failure rate being in inverse proportion to the MTBF in an exponential distribution, the result means that the mean time between failures of the drive gear is shortest and the failure rate of the electricity parts is longest. Figure 3 illustrates the failure rate of the sub-assembly of the tool post and the weak points and failure rate of the main parts. The percentage of each blank indicates the failure rate of each sub-assembly when it is assumed that the failure rate of the tool post is 100%. The Timing Belt, Pulley and Radial Bearing of the Drive Gear, which has the highest failure rate, have a high failure rate and therefore can be expected to be weak parts. As they also have a high failure rate, the Proximity Switch, Quad-Ring (X-Seal), 3 piece type curvic coupling and the Proximity Sensor of the electric parts are expected to break down during actual operation. Figure 4 illustrates the change in the reliability of the tool post and the constituent sub-assembly over time.The sub-assembly that reliability declines sharply is drive gear because the failure rate of the timing belt of the drive gear has a relatively higher failure rate than the other parts. In addition, we found that the reliability of the sub-assembly was almost equal to the failure distribution rate derived from actual customer service data.
3. Manufacture of a reliability tester and reliability testing
3.1 Tool post reliability tester
The failure of a tool post is the failure of indexing and clamping, which are the most important functions of a tool post.This is thought to be the result of a malfunction of the proximity sensor, which senses clamping or leaks caused by wear and tear of the sealing parts. In addition, damage to the main shaft, the defectiveness of parts assembly,the wear of parts due to the repetition of loads, and the backlash caused by loads asymmetry due to biased tool installation also lead to failure. Therefore, the angular resolution, repetition degree, stiffness and flatness of a tool post are very important elements of function and reliability. Table 1 shows the assessment items for a reliability assessment of a tool post. The reference data are made by machine tools maker. Angular resolution and repetition are measured using an angle encoder, and if the values fall outside the reference value, then curvic coupling wear, 0Ring wear and oil pressure decrease are forecast. In the case of wear of the curvic coupling, the stiffness of the tool post decreases. For this measurement, the wear of the curvic coupling can be measured by inflicting loads with a load cell, measuring the transformation value and the stiffness change. Equally, the proximity sensor bracket vibration and temperature increase caused by continual operation can be measured using an accelerometer sensor and thermocouple. In order to measure the aforementioned items, we made a reliability tester for the structure, as shown in Fig. 5. The tester is divided into a drive part, measurement part, control part and supporting part. The drive part is composed of a servomotor to drive the tool post, a hydraulic device and lubricating device; the measured data are processed in the PC. The supporting part is composed of a surface plate on which the tool post reliability tester is installed, and a bracket to which the sensor is fixed, In the study, we also used a surface plate on which a damper is installed.Figure 6 shows the tool post reliability tester which was actually made for this study.
3.2Assessmentofthe performance of the tool post
We measured the performance of the tool post in order to determine the optimum operational conditions for reliability testing. This was conducted so as to define a failure by consecutively measuring performance in a long operation.
3.2.1 Oil pressure and stiffness/repeatability
The stiffness of the radial direction of the tool post is determined by the change of the oil pressure on the curvic coupling. The oil pressure applied in this study is 20~70 kg/em' and the strength inflicted by the load cell is 400 N. In the test, the stiffness decreased noticeably in oil pressure below 40 kg/em' and remained fixed for oil pressure of more than 40 kg/em'. Figure 7 illustrates the change in stiffness according to the change in oil pressure. The optimum stiffness for maintaining the stiffness of a tool post requires oil pressure ofatleast40kg/em'. Moreover, oil pressure also has a great impact on repeatability. If oil pressure is too low, repeatability declines because of the low clamping force of the curvic coupling. However, the use of too high a level of oil pressure is not desirable in the structural aspect either. Figure 8 illustrates the change of repeatability according to oil pressure. Oil pressure proved most desirable at 50 kg/ern" if we consider repeatability;
because repeatability meets the basic value in oil pressure over 50 kg/ern' and fails to meet the basic value in oil pressure below 50 kg/ern',
3.2.2Angularresolution and thermal expansion Angular resolution, for which the absolute basis is difficult to establish, is harder to assess in comparison with repeatability precision. Figure 9 shows the average value of index errors by measuring the angles of each index after incessant operation for 8 hours. As shown in Fig. 9, the basic index is 9, but the indexing error does not show a consistent tendency. Given the offset of the encoder value by the basic index as a result of measurement, we can see an indexing error of about O.03°s. In addition, given that the error does not occur in only one direction, we can see that the error is not caused by the lopsided curvic coupling. We measured the impact by rotating indexes 3 and 9 repeatedly in order to observe the impact according to the angular resolution. Vibration was measured by an acidometer installed on the bracket used to fix the proximity sensor. As Fig. 10 illustrates, we can see that the impact caused by the clamping of the indexingindexing of3isnearly 10times higherthan that ofthe indexing of 9, which supports the findings of the angular resolution test above. If the indexing is conducted incessantly, thermal expansion occurs in the tool post. Thermal expansion can be measured with a gap sensor installed in the radial axial direction. Thermal expansion was found to be 0.5 um in the radial direction and l.Oum in the axial direction after 72 hours of operation. In the following consecutive operation, no more thermal expansion occurred. Therefore, the failure ofatoolpostcanbedefined as follows; (1) oil pressure isbelow 60 kg/ern', (2) the repeatability is over 0.005°, (3) the thermal expansion is over l.Oum in the radial or axial direction, (4) tool post is stopped. That is, one of the mentioned items leadsto failure ofthetoolpost.
4. Reliability assessment
4.1 Reliability test
Through a reliability test, a variety of results may be obtained depending on the test conditions. In this study, on the basis of the performance assessment conducted above, the reliability test was carried out
under oil pressure of 60 kg/crrr', which is thought to keep the stiffuess and repeatability fixed. Although there are other methods, including the eccentric loading, non-loading and consistent loading methods and so forth, for establishing the load condition for the tool post, we adopted non-loading continuous operation for the reliability testing because it was difficult to select the accelerating force. we conducted the operation repeating the rotation and backlash of the tool post index in a sequence of 1~7-->4->10, where the operation time of a cycle was approximately 8 seconds, in order to obtain the test results more quickly. We measured the encoder data and gap sensor data once after 1,000 cycles (that is, measured after 4,000 times of indexing) in order to quantitatively analyze the test results, and conducted 100,000 cycles of consecutive operation. As a result of the reliability test, three failures occurred in total. In the first failure, the tool post stationed itself completely after about 1.9 million cycles with repeatability rapidly falling after about 1.6 million cycles. We checked the oil pressure of the hydraulic motor supplying the clamping force to identify the cause of the failure, but it proved to be working well, as did the proximity switch.
Although the eause of the first failure was not found,the most probable cause seems to be the vibration of the bracket used to fix the proximity switch. The repeatability of the tool post, which started operation after the first halt, was not as high as in the first operation, but we continued the test because it remained within the basic value. In the second failure, the operation stopped after 1.2 million cycles and normal operation resumed after relocation of the bracket fixing the proximity switch, as in the first failure. In this respect, we conducted a deterioration test of the proximity switch, but the, life and performance of the proximity switch were not affected by 6 million on/offs. Therefore the abnormal operation of clamping is sure to be caused by the transformation of the bracket fixing the proximity switch. The tool post stopped at about the millionth cycle after readjustment of the proximity switch.As a result of the analysis of the cause of the failure, the hydraulic motor, which was intended to supply the clamping force, wasn't working because the oil pressure was 0 kg/ern'. Assuming this to be a problem with the hydraulic system, we disassembled the tool post and found, after analysis, that the O-ring, which was intended to transmit the clamping force, had been damaged, and the quad-ring was worn. Figures II and 12 show the damaged part of the O-ring and wear of the quad-ring.
4.2 Reliability analysis
We operated the tool post for about 4.1 million cycles for a reliability test, during which three failures occurred. From this, we reasoned that the life of a tool post is about 1.8 million cycles, that of a hydraulic one is 4.1 million cycles, and that the most frequent cause of failure is not the life of the proximity switch itself but the displacement of the bracket fixing the
proximity switch. We analyzed the reliability of the tool post on the basis of the data obtained from there liability test.The data that should be collected for a reliability analysis include failed parts, failure time, failure mechanism, failure mode, usage conditions and the measures taken against failure, and so forth. Of these, the failure time, usage conditions and number of failures are the important elements to be used in an analysis of the failure rate. That is why we calculated th