tp008多级汽轮机滑动轴承

Optimization of a Leading Edge Groove Tilting Pad Journal Bearing for Application in a Small Optimization of a Leading Edge Groove Tilting Pad Journal Bearing for Application in a Small, High Speed, Multi-Stage Steam Turbine Stephen L. Edney John K. Waite Scan M. Decamillo Dresser-Rand, Wellsville, NY, USA Optimization of a leading edge groove tilting pad journal bearing for application in a small, high speed, multi-stage steam turbine is described. Rotordynamics constraints to meet a design objective maximum operating speed of 18,000 rpm resulted in a rotor with a 5 1. 0 in. bearing span and 5. 0 in. diameter tilting pod journal bearings. This configuration yielded a design with projected bearing loads of less than 25 psi, and journal surface speeds that could approach 400 ft./sec. Under these conditions, the applicable limits of conventional style tilting pad journal bearings are stretched, since operation is well into the turbulent flow regime. This can result in significantly higher than predicted operating pad temperatures and increased frictional losses. Furthermore, at very light pad loads, bearing dynamic performance and influence on rotor behavior often does not correlate well with theory. For this application, high efficiency leading edge groove bearings (journal and thrust) were used, due to their preferred steady state operating characteristics at high speed. However, as is often observed with lightly loaded conventional style journal bearings, dynamic performance did not precisely match that predicted by theory. This was investigated by profiling the exit side of the leading edge groove with both a tapered and, pocket geometry. Two case histories are presented demonstrating their effect on rotor- bearing stability and unbalance response. The modified bearings yielded greater system stability at high speeds, reduced overall vibration amplitudes, and significantly improved effective shaft damping on passing through the rotor's first peak response speed. Introduction The design and optimization of the rotating element and bearings for a small, high speed, multistage steam turbine is described. The design specification was for a steam turbine nominally capable of producing up to 7,000 hp and operating up to a maximum continuous speed of 18,000 rpm. This objective included producing a design that, with minimal hardware change, could be packaged with several combinations of steam and exhaust end assembly, and backpressure to condensing steam path options. This was accomplished with a modular design by fixing the bearing span and standardizing on the steam and exhaust end overhangs. To comply with the dynamics specifications of the American Petroleum Institute (API) Standard 612 [1], the rotor design was optimized by minimizing the bearing span and overhang lengths, and increasing the shaft diameter through the journal bearings. This yielded a design with a bearing span of 5 1.0 in. and 5.0 in. diameter journals that could operate at speeds approaching 400 ft./sec. (at 18,000 rpm) at projected pad loads of less than 25 psi. Under these conditions, bearing operation is well into the turbulent flow regime where pad temperatures and frictional losses can escalate. High pad temperatures can be of concern to OEM's and users of turbomachinery, since temperature limits of the babbitt material can be approached and margin against seizure reduced. Primarily for this reason, directed lubrication bearings were considered essential for this application. Results presented herein are in two sections. The first section compares data from a comprehensive test program, assessing a leading edge groove tilting pad journal bearing with a conventional style flooded design. This evaluation addresses the steady state parameters of pad temperature and power loss, and bearing dynamic influence on rotor response. The second section discusses the application and operational performance of a leading edge groove journal bearing in three production applications. Two case histories are reviewed of a modification to the bearing that greatly improved dynamic performance in this light load application. Bearing Evaluation The requirements for new bearing designs are clear: increased load carrying capacity and lower frictional losses at reduced oil flowrates, without. increasing operating pad temperature. In pursuit of this, bearing manufacturers initially focussed their efforts on thrust bearing design [2). These improved designs first utilized the concept of directed lubrication technology, and have been widely used on OEM machinery for a little over a decade. Thrust bearings, by far, consume the most power of the total bearing system, and were the logical starting point for the development of this technology. However, it is only in recent years that this same technology has been actively applied to tilting pad journal bearings [3, 4, 5, 6]. This is because machine speeds are now nearing the limits of acceptable performance of conventional style journal bearing designs [7], particularly when service conditions require operation into the turbulent flow regime. Under these conditions, bearing losses exponentially increase, which can have a noticeable effect on a machine's overall efficiency. Likewise, increasing the oil flowrate to reduce high bearing temperatures generally only increases the size and cost of the lubrication system, and often with minimal or no effect. Prior to the first application of a leading edge groove journal bearing in a production turbine, tests were file:///F|/USERS/mafisher/MYDOCS/TechPapers/tp008/tp008prt.htm (1 of 36) [07/29/2002 3:57:44 PM] Optimization of a Leading Edge Groove Tilting Pad Journal Bearing for Application in a Small conducted in a laboratory vehicle to compare the bearing's performance against a conventional style flooded design. The purpose of this test was twofold. First, to verify the claimed improvement in steady state (temperature and power loss) performance, and, second, to ensure of no adverse dynamic (stiffness and damping) effects on rotor behavior. Test Vehicle The test vehicle was a laboratory turbine capable of operating up to a maximum speed of 16,500 rpm. A general arrangement of the test vehicle is shown in Figure 1. The rotor was of an integrally forged design with five wheels and a bearing span of 59.98 in., a midspan shaft diameter of 6.0 in., and 4.0 in. diameter journals. The first wheel (control stage) only was bladed and steam provided the motive driving force. All other wheels were blank and were for the purpose of simulating a multi-stage rotor. The rotor is shown with shaft end flanges and bored overhangs that can accept different weights for the purpose of simulating overhung moments. Test Bearings The test bearings were of 4.0 in. diameter, 0.75 length/diameter ratio (L/D), and four pad configuration. The diametral assembled bearing clearance was nominally 1.5 mil./in. of journal diameter, and loading was directed between the two lower half pads. Both bearings were designed with pad axial and circumferential alignment capability. The projected pad loads were 34.9 psi and 39.9 psi at the steam and exhaust end bearing locations, respectively. Specific features relevant to each of the test bearings are described below. Pertinent details are summarized in Table 1. Values for the assembled diametral bearing clearance were obtained by direct measurement using a precision vertical mandrel, horizontal table and dial indicator. Pad preload was based on the nominal design pad bore and measured bearing clearance. The lubricant was a light turbine oil (equivalent to ISO VG 32) with a viscosity of 150 SSU at 100ºF. Flooded Bearing A sectional view of the flooded design tested is shown in Figure 2. The bearing was designed with separate oil supply and oil drain orifice plugs to control flowrate. These were located in the bearing housing between each pair of neighboring pads. Oil supply was provided by one orifice placed at the axial midposition. Oil drain was controlled via two orifices, one placed near each axial edge of the pad. Their respective locations are identified on Figure 2. Tight clearance floating ring end seals were installed at each end of the housing to ensure a flooded bearing. file:///F|/USERS/mafisher/MYDOCS/TechPapers/tp008/tp008prt.htm (2 of 36) [07/29/2002 3:57:44 PM] Optimization of a Leading Edge Groove Tilting Pad Journal Bearing for Application in a Small Leading Edge Groove Bearing A sectional view of the leading edge groove design is shown in Figure 3. Oil supply to the bearing was from an annulus on the outside of the housing through a single feed tube to each pad. These feed tubes supply cool oil directly to the center of a groove machined axially across the leading edge of each pad. This groove essentially is an extension of the pad that provides a plenum of cool oil that directly feeds the bearing. The resultant cooler oil supplied to the oil film insulates the babbitt face from hot oil carryover that file:///F|/USERS/mafisher/MYDOCS/TechPapers/tp008/tp008prt.htm (3 of 36) [07/29/2002 3:57:44 PM] Optimization of a Leading Edge Groove Tilting Pad Journal Bearing for Application in a Small adheres to the rotating journal, thus greatly reducing the babbitt temperature. Consequently, the lower pad temperatures permit lower oil flow requirements which contribute to a reduction in bearing power loss [8]. This bearing was designed with fixed open clearance end seals to allow spent hot oil to exit axially without restriction and contamination of the fresh cool supply oil. Radial drain slots located near bottom dead center also were provided in each seal. Instrumentation The two lower half (statically loaded) pads only were instrumented, each with a single type 'J' thermocouple temperature sensor. The sensors were located at the 75 percent position from the pad leading edge on the axial centerline, in accordance with industry standard guidelines given in API 670 [9]. This location is generally regarded as corresponding to the vicinity of the critical pad temperature. The sensors were embedded in the pads to a depth of 50 to 60 thousandths of an inch below the babbitt bond line. The babbitt material was an additional 50 to 60 thousandths of an inch thick. Thermocouple temperature sensors also were used to measure the oil supply and oil drain temperatures. Turbine type flowmeters were used to measure the oil flowrate. The temperature rise across the bearing and the oil flowrate were used to derive a value for power loss. This was possible only on the exhaust end journal bearing. The steam end journal bearing was located in the same bearing case as the thrust, with a common oil supply and oil drain line. Radial vibration was measured using four displacement proximity probes; two located just inboard of each journal bearing 45 degrees either side of top dead center. Shaft speed was measured from a 30 tooth wheel and displayed on a digital counter. Test Procedure Testing was divided into two parts separately covering an evaluation of steady state performance, and of dynamic influence on rotor response. The steady state tests were conducted over a speed range from 4,000 to 16,000 rpm. The turbine was assembled with the relevant test bearings. The rotor was then run up to a speed of 4,000 rpm. This speed was maintained until conditions stabilized. With the oil supply temperature regulated to 120°F at a pressure of 20 psig, shaft speed was increased to 16,000 rpm in increments of 1,000 rpm. At each increment, two sets of readings were taken; one immediately following a speed increment, and one after pad temperature had stabilized. Data recorded were oil pressure, oil flowrate, and temperature. Temperature measurements were of the four instrumented pods, oil supply, and oil drain. Bearing dynamic characteristics were evaluated by comparing the rotor's response to a first mode unbalance excitation. The unbalance weight used was equivalent to 4.0 oz.- in. placed at the rotor's midspan trim balance plane. With the weight installed, the rotor was run up to a speed of 16,500 rpm and then allowed to coast down in the absence of steam forces. During coast down, vibration data was acquired at all four radial probes over a speed range from 16,000 to 500 rpm. file:///F|/USERS/mafisher/MYDOCS/TechPapers/tp008/tp008prt.htm (4 of 36) [07/29/2002 3:57:44 PM] Optimization of a Leading Edge Groove Tilting Pad Journal Bearing for Application in a Small Test Results Evaluation of steady state performance was in terms of pad temperature and power loss, and dynamic behavior in terms of rotor response to an unbalance excitation. Since power loss could only be derived at the exhaust end bearing, all of the results presented are for this location only. Table 2 summarizes the measured pad metal temperatures from each test, and Table 3 compares power loss over a speed range of between 9,000 and 16,000 rpm. The following notation applies: 'EE' refers to exhaust end bearing, and 'left' and 'right' to the thermocouple and proximity probe locations viewed from the turbine governor. When referring to thermocouple location, left is the sensor located in the downstream pad and right is the sensor located in the upstream pad. file:///F|/USERS/mafisher/MYDOCS/TechPapers/tp008/tp008prt.htm (5 of 36) [07/29/2002 3:57:44 PM] Optimization of a Leading Edge Groove Tilting Pad Journal Bearing for Application in a Small file:///F|/USERS/mafisher/MYDOCS/TechPapers/tp008/tp008prt.htm (6 of 36) [07/29/2002 3:57:44 PM] Optimization of a Leading Edge Groove Tilting Pad Journal Bearing for Application in a Small Flooded Bearing Baseline steady state data of the flooded bearings are given in Figures 4 and 5. Both sets of temperature data increase with speed up to about 14,000 rpm. Beyond this speed, there is some indication that the temperature may have reached a plateau and started to decrease. The individual pad temperatures exhibit some divergence, however, with the peak value reaching 234°F at the pad located in the upstream position. This is contradictory to the generally accepted view that, with a load between pad arrangement, the file:///F|/USERS/mafisher/MYDOCS/TechPapers/tp008/tp008prt.htm (7 of 36) [07/29/2002 3:57:44 PM] Optimization of a Leading Edge Groove Tilting Pad Journal Bearing for Application in a Small pad located in the downstream position will indicate a higher trailing edge temperature due to mixing of hot oil carried over from the upstream pad. However, this is consistent with results from a series of tests carried out on bearings with similar pad loadings [7], and from tests with considerably higher pad loadings [10]. In contrast, the power consumed by the bearing continually increases with speed reaching a value of 27.5 hp at 16,000 rpm. There is a clear increase in rate of power loss at around 12,000 rpm. This characteristic is generally regarded as indicative of the region of transition from laminar to turbulent flow. Vibration data is given in Figure 6 for a first mode excitation (midspan unbalance). On passing through the first critical speed, three response peaks are clearly evident at the probes located on the left side, and two response peaks at the probes located on the right side. Leading Edge Groove Bearing Steady state data of the leading edge groove bearings are given in Figures 7 and 8. Clearly, the temperature data monotonically increases with speed with only a slight leveling off at around 7,000 rpm. At 16,000 rpm, the maximum temperature reached is 186°Fat the pad located in the upstream position. The downstream pad indicates a slightly lower temperatures of between 5-10°F. A reduction in pad temperature of some 35-45°F compared with the flooded design is apparent, and with 50 percent less oil flowrate. This is partially reflected in power loss, particularly at the higher end of the speed range, which reaches a value of only 20.3 hp. At 16,000 rpm, this represents a reduction of approximately 25 percent compared with the flooded design. Vibration data with the same unbalance condition is given in Figure 9. These results exhibit very similar vibration trends compared to the flooded design. The amplitude and location of discrete response peaks and troughs are near identical, as are the overall vibration levels. Clearly, the leading edge groove feature of this design has no adverse effects on bearing dynamic performance at the speed and load conditions examined. file:///F|/USERS/mafisher/MYDOCS/TechPapers/tp008/tp008prt.htm (8 of 36) [07/29/2002 3:57:44 PM] Optimization of a Leading Edge Groove Tilting Pad Journal Bearing for Application in a Small file:///F|/USERS/mafisher/MYDOCS/TechPapers/tp008/tp008prt.htm (9 of 36) [07/29/2002 3:57:44 PM] Optimization of a Leading Edge Groove Tilting Pad Journal Bearing for Application in a Small High Speed, Multi-Stage Steam Turbine The purpose of testing the leading edge groove bearing was for application in a small, high speed, multi-stage steam turbine. design specifications were for a turbine nominally capable of producing up to 7,000 hp and operating up to a maximum speed of 18,000 rpm. An objective of this design was to produce, with minimal hardware change, a turbine that could be built with several combinations of steam and exhaust end assembly, and backpressure to condensing steam path options. This was achieved with a modular approach by fixing the bearing span and standardizing on the steam and exhaust end overhangs. The final design has available several single and multi valve steam end options, and backpressure to condensing steam paths of up to five stages. A typical configuration is illustrated in Figure 10 which depicts a single valve, five stage condensing design. file:///F|/USERS/mafisher/MYDOCS/TechPapers/tp008/tp008prt.htm (10 of 36) [07/29/2002 3:57:44 PM] Optimization of a Leading Edge Groove Tilting Pad Journal Bearing for Application in a Small Rotor Design A schematic of a typical five stage condensing rotor is shown in Figure I I . To achieve the objective operating speed of 18,000 rpm in compliance with current rotordynamics specifications [1], the bearing span and overhangs were kept to a minimum, and the shaft diameter through the journal bearings increased. This yielded a base design with a 51.0 in bearing span and 5.0 in. diameter journals. The midshaft diameter was left variable in order that the rotor's dynamic characteristics could be tuned in relation to the operating speed range. For example, a five stage rotor operating at speeds approaching 18,000 rpm would require a midshaft diameter of 8.5 in., to raise the location of the third lateral peak response speed to attain the required separation margin. At lower operating speeds, the midshaft diameter can be reduced to 6.5 in. to allow operation between critically damped first and second peak response speeds. To minimize the steam end overhang moment, speed pickup teeth were machined on the outside diameter of the thrust collar to eliminate a toothed wheel. The thrust bearing was kept as small as possible by stepping down the shaft diameter outboard of the journal. A flanged face was provided for axial displacement probes, with grounding brushes located near the axial centerline. The exhaust end overhang was minimized by tucking the bearing case under the diffuser portion of the exhaust, and making the coupling flange integral with the rotor. Provision was made for two radial displacement probes per end, located inboard of each journal bearing 45 degrees either side of top dead center. Two trim balance planes were provided as standard, one located at each end of the central shaft section. Bearing Design file:///F|/USERS/mafisher/MYDOCS/TechPapers/tp008/tp008prt.htm (11 of 36) [07/29/200