Stability Characteristics of Supercritical High-Pressure Turbines Depending on the Designs of Tilting Pad Journal Bearings

Abstract

In this study, for a high-pressure turbine (HPT) of 800 MW class supercritical thermal-power plant, considering aerodynamic cross-coupling, we performed a rotordynamic logarithmic decrement (LogDec) stability analysis with various tilting pad journal bearing (TPJB) designs, which several steam turbine OEMs (original equipment manufacturers) currently apply in their supercritical and ultra-supercritical HPTs. We considered the following TPJB designs: 6-Pad load on pad (LOP)/load between pad (LBP), 5-Pad LOP/LBP, Hybrid 3-Pad LOP (lower 3-Pad tilting and upper 1-Pad fixed), and 5-Pad LBPs with the design variables of offset and preload. We used the API Level-I method for a LogDec stability analysis. Following results are summarized only in a standpoint of LogDec stability. The Hybrid 3-Pad LOP TPJBs most excellently outperform all the other TPJBs over nearly a full range of cross-coupled stiffness. In a high range of cross-coupled stiffness, both the 6-Pad LOP and 5-Pad LOP TPJBs may be recommended as a practical conservative bearing design approach for enhancing a rotordynamic stability of the HPT. As expected, in a high range of cross-coupled stiffness, the 6-Pad LBP TPJBs exhibit a better performance than the 5-Pad LBP TPJBs. However, contrary to one's expectation, notably, the 5-Pad LOP TPJBs exhibit a slightly better performance than the 6-Pad LOP TPJBs. Furthermore, we do not recommend any TPJB design efforts of either increasing a pad offset from 0.5 or a pad preload from 0 for the HPT in a standpoint of stability.

1. Introduction

In worldwide coal-fired thermal-power plants have been progressing continuously for both efficiency and capacity improvements from the subcritical (efficiency: 35%), the supercritical (SC, efficiency: 38%), and to the ultra-supercritical (USC, efficiency: 42% or higher, e.g., 49%) power plants. In South Korea, also, following this worldwide technical trend, the subcritical (Samcheonpo # 1~4 units: 560 MW, 538^oC, and 16.5 MPa), the Korean 500 MW standard semi-SC (Samcheonpo # 5, 6 units: 500 MW, 538°C, and 24.8 MPa), the SC (Yeongheung #1, 2 units: 800 MW, 566^oC, and 24.8 MPa), and the USC (New Boryeong # 1, 2 units: 1,000 MW, 610^oC, and 26.0 MPa) power plants have been commercialized in turn.

Steam turbine trains of large capacity coal-fired power plants consist of HIP (high and intermediate-pressure) turbines for the semi-SC condition, and HP (high-pressure) and IP (intermediate-pressure) turbines, separately, for the SC and USC conditions, together with LP (lowpressure) turbines. A rotor weight of typical HIP or HP turbine is in a range of 15 to 20 ton. Since a prevention of rotordynamic instability due to steam whirl is a toppriority design issue because of high steam temperature and pressure conditions, several global OEMs have been adopting 6-Pad LOP TPJBs as support bearings for these rotors. This practice is more noticeable in bearing designs for the SC and USC application HPTs.

TPJBs require a high-level of design engineering. Number of pads, load support type (LOP, LBP), offset, preload, and bearing clearances etc. are key design variables[1-6]. These design variables affect not only operating temperature characteristics of TPJBs but also their fluid film stiffness and damping coefficients, which are closely related to rotor vibration characteristics, e.g., stability. Nicholas et al.[7] reported that the TPJB designs with zero preload, center pivot (or zero offset), and LOP provide the most stable rotordynamic characteristics for high-speed compressors. Lee [8] carried out a rotordynamic stability analysis of a light-weight (rotor mass: 240 kg) and high-speed (9,540 rpm) process 8- stage centrifugal compressor, where a design of decreasing the TPJB's preload improved the stability whereas there was no stability difference between LOP and LBP because the rotor was of light-weight and operated in high-speed. Ikeno et al.[9] investigated the design effect of large (journal diameter: 320~400 mm) TPJBs for mega ethylene plant applications. Among their reviewed 4-Pad LBP, 5-Pad LBP, and 5-Pad LOP TPJBs, 5-Pad LOP TPJBs were the best in a standpoint of rotordynamic stability considering gas-induced cross-coupling, i.e., aerodynamic cross-coupled stiffness forces by impellers and seals etc. Zeidan[2] explained that when a gas-induced cross-coupling is acting, the reason why the LOP designs are better than the LBP ones in a standpoint of the rotordynamic stability is that as in case of the LBP TPJBs the shaft orbit is circular whilst in case of the LOP TPJBs the shaft orbit is elliptical, the LOP TPJBs have less instability excitation energy than the LBP TPJBs, expressed as a product of orbit area and cross-coupled stiffness, A (Kxy-Kyx). API STD 617[10] and 684[11] describe a method of evaluating the dynamic stability of rotor system by calculating LogDec, depending on applied cross-coupling.

As a groundwork Lee and Jang[12] performed a lubrication performance analysis of TPJBs, applied to a HPT. In this study, for the HPT of Yeongheung # 1, 2 units of 800 MW class SC power plant, considering aerodynamic cross-coupling by high-temperature and highpressure steam, we performed a rotordynamic LogDec stability analysis with not only originally applied 6- Pad LOP TPJB designs but also other various TPJB designs which several steam turbines OEMs currently use in their SC or USC HPTs, and compared results with each other. Specifically, the considered TPJB designs were 6-Pad LOP/LBP, 5-Pad LOP/LBP, Hybrid 3-Pad LOP (lower 3-Pad tilting and upper 1-Pad fixed), and 5-Pad LBPs with design variables of offset and preload.

2. Bearing and Rotor Analysis Models

In Table 1 are given the design and operation condition data of HPT TPJBs in 800 MW class SC turbine train. Figure 1 represents a lubrication analysis model of HPT 6-Pad LOP #1 TPJB for offset = 0.5 and m = 0.0.

Table 1. Design and operation condition data of HPT TPJBs

Fig. 1. Lubrication analysis geometry model of HPT 6- Pad LOP #1 TPJB for offset = 0.5 and m = 0.0.

In difference with any usual FE rotordynamic analysis rotor model, now, for a FE rotordynamic analysis with a consideration of cross-coupling, Fig. 2 represents an applied LogDec, δ_A, calculation-purpose rotor model, having an aerodynamic cross-coupling element at the station # 23 in the middle of the rotor. Figure 2 also shows an example whose 1st eigenvalue analysis results give a whirl natural frequency of 2,470 rpm and δ_A = 0.387 for a cross-coupled stiffness, δ_A = 1.5e + 08 N/m, acting at 3,600 rpm.

Fig. 2. A FE rotordynamic analysis model of HPT rotor with a cross-coupling element added in the middle and its 1st mode shapes in red, simultaneously, in horizontal and vertical planes at 3,600 rpm with 6-Pad LOP TPJBs: m = 0.0 and offset = 0.5, for Q_A = 1.5e + 08 N/m.

3. Results and Discussion of Applied LogDec Stability Calculation Depending on TPJB Designs

A LogDec stability analysis of the HPT rotor was carried out with applying a total cross-coupled stiffness Q_A, acting on all multi-stage blades by steam, at the middle stage position. As Q_A increased from 0 to 2.8e + 08 N/m (which was wide enough to show its overall effect on LogDec within a range of ±0.8), a change of LogDec δ_A associated with a specific TPJBs design was calculated. Here, all applied LogDec analyses were performed only for the 1st whirl mode at 3,600 rpm, utilizing the FE rotor model shown in Fig. 2.

3-1. LogDec Effects of LOP and LBP in 6-Pad TPJBs

For 6-Pad TPJBs: m = 0.0 and offset = 0.5, depending on Q_A, Fig. 3 shows in comparison the δ_A characteristics of LOP and LBP designs. In a standpoint of stability, in a low range of 0 < Q_A < 1.36e + 08 N/m the LBP performs better than the LOP whereas in a high range of Q_A > 1.36e + 08 N/m the LOP greatly outperforms the LBP. That is, as a conservative bearing design against a possible aerodynamic cross-coupling instability induced by high temperature and pressure steam in the HPT, the 6-Pad LOP TPJBs shall be recommended over the 6-Pad LBP TPJBs.

Fig. 3. Applied LogDec characteristics, at 3,600 rpm, of the HPT rotor with 6-Pad TPJBs, depending on the LOP and LBP: m = 0.0 and offset = 0.5.

3-2. LogDec Effects of LOP and LBP in 5-Pad TPJBs

For 5-Pad TPJBs: m = 0.0 and offset = 0.5, depending on Q_A, Fig. 4 shows in comparison the δ_A characteristics of LOP and LBP designs. In a standpoint of stability, similarly, in a low range of 0 < Q_A < 1.13e + 08 N/m the LBP performs better than the LOP whereas in a high range of Q_A >1.13e+08N/m the LOP greatly outperforms the LBP. That is, as a conservative bearing design against a possible instability by high temperature and pressure steam in the HPT, the 5-Pad LOP TPJBs shall be recommended over the 5-Pad LBP TPJBs.

Fig. 4. Applied LogDec characteristics, at 3,600 rpm, of the HPT rotor with 5-Pad TPJBs, depending on the LOP and LBP: m = 0.0 and offset = 0.5.

3-3. LogDec Effects of 6-Pad and 5-Pad in LBP TPJBs

For LBP TPJBs: m = 0.0 and offset = 0.5, depending on Q_A, Fig. 5 shows in comparison the δ_A characteristics of 6-Pad and 5-Pad designs. In a standpoint of stability, in a low range of 0 < Q_A < 1.13e + 08 N/m the 5-Pad performs better than the 6-Pad whereas in a high range of Q_A > 1.13e + 08 N/m the 6-Pad performs better than the 5-Pad. That is, as a conservative bearing design against a possible instability by high temperature and pressure steam in the HPT, the 6-Pad LBP TPJBs shall be recommended over the 5-Pad LBP TPJBs as expected.

Fig. 5. Applied LogDec characteristics, at 3,600 rpm, of the HPT rotor with LBP TPJBs, depending on the 6-Pad and 5-Pad: m = 0.0 and offset = 0.5.

3-4. LogDec Effects of 6-Pad and 5-Pad in LOP TPJBs

For LOP TPJBs : m = 0.0 and offset = 0.5, depending on Q_A, Fig. 6 shows in comparison the δ_A characteristics of 6-Pad and 5-Pad designs. In a standpoint of stability, in a full range of Q_A the 5-Pad performs slightly better than the 6-Pad. Therefore, as a conservative bearing design against a possible instability by high temperature and pressure steam in the HPT, the 5-Pad LOP TPJBs shall be recommended over the 6-Pad LOP TPJBs, where this is contrary to one's expectation.

Fig. 6. Applied LogDec characteristics, at 3,600 rpm, of the HPT rotor with LOP TPJBs, depending on the 6-Pad and 5-Pad: m = 0.0 and offset = 0.5.

3-5. LogDec Effects of Offset and Preload in 5-Pad LBP TPJBs

For 5-Pad LBP TPJBs, the effects of offset and preload, m, which are key design variables of TPJBs, were analyzed.

For a fixed m = 0.0, depending on Q_A, Fig. 7 shows in comparison the δ_A characteristics as an offset increases from 0.5 and to 0.55 and 0.6. It is observed that over a full range of Q_A, δ_A decreases slightly as an offset increases from a center pivot of 0.5 to 0.55 whereas δ_A decreases greatly with an offset = 0.6. Therefore, TPJBs design with an offset = up to 0.55 may be considered to decrease bearing temperature and increase stiffness but TPJBs design with an offset = 0.6 shall not be recommended for the HPT, in a standpoint of LogDec, i.e., stability.

Fig. 7. Applied LogDec characteristics, at 3,600 rpm, of the HPT rotor with 5-Pad LBP TPJBs, depending on the offset: m = 0.0 and offset = 0.5, 0.55, 0.6.

For a fixed offset = 0.5, depending on Q_A, Fig. 8 shows in comparison the δ_A characteristics as m increases from 0.0 and to 0.2 and 0.4. It is observed that in a low range of 0 < δ_A < 1.0e+ 08N/m δ_A decreases nearly equally as m increases to 0.2 and 0.4 whereas in a high range of Q_A > 1.0e + 08 N/m δ_A decreases in a greater magnitude with m = 0.4 than with m= 0.2.

Fig. 8. Applied LogDec characteristics, at 3,600 rpm, of the HPT rotor with 5-Pad LBP TPJBs, depending on the preload: m = 0.0, 0.2, 0.4 and offset = 0.5.

Showing the results of Figs. 7 and 8, together, in superposition, depending on Q_A, Fig. 9 represents the δ_A characteristics as an offset increases from 0.5 and to 0.55 and 0.6, for a fixed m =0.0 and as m increases from 0.0 to 0.2 and 0.4, for a fixed offset = 0.5. Particularly, in a range of Q_A > 1.0e + 08 N/m the effects that an offset independently increases to 0.55 and 0.6 and that m independently increases to 0.2 and 0.4 are almost the same on δ_A.

Fig. 9. Applied LogDec characteristics, at 3,600 rpm, of the HPT rotor with 5-Pad LBP TPJBs, depending on the offset and preload, independently: m = 0.0, 0.2, 0.4 and offset = 0.5, 0.55, 0.6.

Specially, for a simultaneous application of offset = 0.6 and m = 0.4, Fig. 10 shows in comparison the δ_A characteristics, depending on Q_A. It is observed that for offset = 0.6 and m = 0.4, δ_A decreases in a greater magnitude over a full range of Q_A. Therefore, it is reviewed that a simultaneous design application of offset = 0.6 and m = 0.4 shall be quite undesirable in a standpoint of stability.

Fig. 10. Applied LogDec characteristics, at 3,600 rpm, of the HPT rotor with 5-Pad LBP TPJBs, depending on the offset and preload, independently (m = 0.0, 0.4 and offset = 0.5, 0.55) and simultaneously (m = 0.4 and offset = 0.6).

3-6. Comprehensive LogDec Effects of various TPJB Designs

Depending on Q_A, Fig. 11 shows in a comprehensive comparison the δ_A characteristics of various TPJB designs, i.e., 6-Pad LOP, 6-Pad LBP, 5-Pad LOP, 5-Pad LBP, Hybrid 3-Pad LOP (lower 3-Pad tilting and upper 1- Pad fixed), and 5-Pad LBPs with design variables of offset and preload.

Fig. 11. Applied LogDec characteristics, at 3,600 rpm, of the HPT rotor with various TPJBs designs in summary.

It is observed that in a standpoint of stability the Hybrid 3-Pad LOP TPJBs most excellently outperform all the other TPJBs over nearly a full range of aerodynamic cross-coupled stiffness, Q_A. Also, in a high range of Q_A > 1.2e + 08 N/m, as a practical conservative bearing design approach for enhancing a rotordynamic stability of the HPT both the 6-Pad LOP and 5-Pad LOP TPJBs may be recommended as well.

4. Conclusions

For a HPT of 800 MW class supercritical thermalpower plant, considering aerodynamic cross-coupling induced by high-temperature and high-pressure steam, we performed a rotordynamic LogDec stability analysis with applying various TPJB designs, which several steam turbine OEMs currently use in their supercritical or ultra-supercritical turbines. We considered the following TPJB designs: 6-Pad LOP, 6-Pad LBP, 5-Pad LOP, 5- Pad LBP, Hybrid 3-Pad LOP (lower 3-Pad tilting and upper 1-Pad fixed), and 5-Pad LBPs with design variables of offset and preload.

Following results are summarized only in a standpoint of LogDec stability. The Hybrid 3-Pad LOP TPJBs most excellently outperform all the other TPJBs over nearly a full range of cross-coupled stiffness. Besides, in a high range of cross-coupled stiffness, both the 6- Pad LOP and 5-Pad LOP TPJBs may be recommended as well as a practical conservative bearing design approach for enhancing a rotordynamic stability of the HPT. As expected, in a high range of cross-coupled stiffness the 6-Pad LBP TPJBs exhibit a better performance than the 5-Pad LBP TPJBs. However, in contrast to one's expectation, notably, the 5-Pad LOP TPJBs exhibit a slightly better performance than the 6-Pad LOP TPJBs over a full range of cross-coupled stiffness. Furthermore, we do not recommend any TPJBs design efforts of either increasing a pad offset from 0.5 or a pad preload from 0 for the HPT, in a standpoint of stability.