外文翻譯--研究一個離心泵內(nèi)的震動壓力—利用數(shù)字表示的方法來支持工業(yè)實驗

1、<p><b>　　英文文獻</b></p><p>　　Investigation into pressure pulsations in a centrifugal pump using numerical methods supported by industrial tests</p><p>　　R.Spence and J.Amaral-Teix

2、eira </p><p><b>　　ABSTRACT </b></p><p>　　The operation of centrifugal pumps can generate instabilities and pressure pulsations that may be detrimental to the integrity and performanc

3、e of the pump. In the present study a numerical investigation of the time variation of pressure within a complete pump was undertaken. A range of parameters and three flow rates were investigated and the pulsations were

4、extracted at 15 different locations covering important pump locations experiencing the largest pulsation levels. It was also note that moni</p><p>　　Introduction </p><p>　　Centrifugal pumps have

5、 been developed and refined over many years. In practice the design of both the impeller and volute are complex, with numerous geometrical parameters being required to identify a design that will form a hydraulically eff

6、icient pump. Even with tried designs it is well known that the operation of rotodynamic pump can result in the generation of the design process used; the final decision regarding the suitability of any significantly new

7、pump design is usually made following </p><p>　　Typically these investigations are conducted with a view to reducing, or eliminating the number of tests conducted and to highlight any undesirable design char

8、acteristics at an early stage. Although commercial CFD packages have been used to predict time dependent pressure pulsations, computational facilities seem to have limited most of that work to simulating the volute and i

9、mpeller interactions only, without the suction inlet branch and leakage flow paths being considered. The current work aim</p><p>　　The numerical model incorporates all of the major flow paths in a pump encom

10、passing the suction inlet, impeller, leakage pathways and the volute casing. The work focuses on a reduced scale version of a high energy impeller in a double entry, single stage pump arrangement. The understanding of pr

11、essure pulsations by pump manufacturers seems surprisingly limited. No official standards exist for safe levels of pressure pulsations in pumps; the only industry adopted guideline is a guarantee of less </p><

12、p>　　The analyses presented here are part of a large parametric study investigating the effect of internal geometry on pressure pulsations within the pump. Due to the constrains involved in such a large study, one addi

13、tional aim of this work was to show that a reasonable estimate of pressure pulsations and trends in pressure pulsations can be achieved, for differing pump geometries, in a reasonable time frame by numerical means.</p

14、><p>　　Experimental Investigation </p><p>　　Some experimental results were available to one of the authors from industrial tests performed to examine pressure pulsations in a reduced scale contract

15、 pump. This experimental work performed a number of years before the numerical analyses. The pump received fluid from a closed system, such that the pump was situated with a bend 3.5 diameters upstream and a second bend

16、4 diameters downstream. Directly following the upstream bend, flow straighteners were fitten in order to reduce the flow effect</p><p>　　One Z type Entran pressure transducer was utilized in the experimental

17、 tests. This transducer was mounted in the impeller shroud 15omm behind the suction face of an impeller blade. The electrical signals from the transducers were transferred from the rotating element the stationary data re

18、corder via a slip ring arrangement at the non-drive end of the pump. The test ring was run at an initial speed of 1400RPM over the flow range 25-125%of duty with 5-min tape recordings of the transducers being t</p>

19、<p>　　Due to the time constraints involved in the experimental tests, detailed performance data was not taken. The test work was conducted in accordance with normal industrial practices and it is considered that t

20、he pressures were accurate within 1.5%.</p><p>　　Numerical Investigation </p><p>　　CFX-TASC flow is a commercial 3D-CFD code that a finite-element based finite-volume method to the transport equ

21、ations. This method provides the benefit of retaining the geometric flexibility of finite-element methods while retaining the conservation properties of the finite-volume method, I, e low numerical error on non-smooth gi

22、rds. The code employs co-located structured grids and a coupied algebraic multigrid scheme to solve the system of equations. It is a fully-implicit solver, thus it creat</p><p>　　3、1 Model generation </p&

23、gt;<p>　　The impeller geometry was created using CFX-Bladehen as two mirrored halves, using the maximum diameter of 366omm. The bull-nose aspect of the inline impeller design proved difficult to capture within CFX

24、-Bladegen, as the program is not designed to model double entry impeller and does not allow for hub profiles that terminate in the bull nose arrangement. To avoid this problem, although the true hub effectively finished

25、at the point of symmetry, a fictitious hub was extended from the midpoint of</p><p>　　CFX-Turbogrid was used to generate grid around the blades on either side of the impeller, with the grid on both sides of

26、the impellers being identical. For the inline impeller, care was taken to ensure that the grid was radial at the position of the bull nose. This is essential to aid the creation of an interface between the impellers in C

27、FX-TASC flow, however it also limited the quality of the grid. With an initial impeller grid estimate of around 500,000 modes, it is immediately apparent that </p><p>　　4、2 Solution parameters</p><

28、;p>　　As the motion of the impeller blades relative to the stationary volute is central to the investigation, the analysis must involve multiple frames of reference. In order to prevent two rotor/stator interfaces bein

29、g perpendicular to each other at the ompeller outlet, the leakage flow paths and the impeller grids were set in the rotational frame of reference. The suction inlet and volute grids were in the stationary frame of refere

30、nce. The grid interfaces used in the study can be summarized as foll</p><p>　　Internal component connection : general grid interface(GGI)</p><p>　　Between two stationary components: general grid

31、 interface</p><p>　　Between two rotational components: general grid interface</p><p>　　Between rotational and stationary components:</p><p>　　Frozen rotor interface (steady state an

32、alysis)</p><p>　　Rotor/stator interface (transient analysis)</p><p>　　The frozen rotor interface achieves frame change without relative position change over time and without interfacial averagin

33、g. Local flow features are allowed to propagate. A physical analogy is to imagine observing the flow between a stationary and rotating component under stroboscopic lighting. The rotor/stator interface is a sliding(frame

34、change) interface that can account for transient interaction effects. The mesh on either side of a rotor/stator interface is always in relative motion with </p><p>　　Inlet: total pressure “Outlet: mass flow”

35、</p><p>　　Inlet: mass flow “Outlet: static pressure”</p><p>　　Inlet: total pressure “Outlet: static ptessure”</p><p>　　The transient analysis was conducted for three flow rates. The

36、se are defined as being the duty flow condition (1.00oQn), 0.50oQn and 0.25oQn. Turbulence has been simulated using the standatdk-epsilon model. An investigation was conducted into various turbulence models and the stand

37、ard k-epsilon model was found to be more stable and produce better results than other models over the flow range being considered. The grid size used in the analysis is not capable of modeling local boundary layers hen&l

38、t;/p><p><b>　　中文翻譯</b></p><p>　　研究一個離心泵內(nèi)的震動壓力—利用數(shù)字表示的方法來支持工業(yè)實驗</p><p><b>　　摘要</b></p><p>　　離心泵的操作會產(chǎn)生不穩(wěn)定性和震動壓力，這可能有害于泵的完整性和其他性能。當前，正在研究一個承受于整個離心泵內(nèi)部，

39、隨時間變化的壓力。計算出參數(shù)的范圍和三個流動速度，并且在泵的15個被覆蓋不同的重要地方萃取震動。瞬變流動導致在一個小范圍試驗測量所獲取的結(jié)果與明顯的顯示泵的位置經(jīng)歷大的震動水平相比較。這也說明，在泵渦形保護套的上死點對泵內(nèi)在震動的保護要比在流動中更好。</p><p><b>　　1、介紹</b></p><p>　　離心泵早已被發(fā)明，并且精確度超過了很多年前。在實踐

40、中，葉輪和渦螺的設計是一個整體，利用大量必要的幾何學參數(shù)驗證這個設計這將會產(chǎn)生一個水利的效率泵。眾所周知，用試驗就能導致操作動態(tài)泵時產(chǎn)生震動壓力。</p><p>　　不管習慣的設計過程，任何重要的新泵設計最后能相匹配的結(jié)果通常都是跟隨物理實驗。這些實驗在時間和資金上常常都有較大的花費。例如，由于制造業(yè)的模型裝置，原裝泵也可以裝配和利用實驗設備。逐漸地，泵制造廠商把計算法轉(zhuǎn)向機構(gòu)特點。</p>&l

41、t;p>　　引導這些與代表性的研究著眼于減少體重法，或者排除實驗中的數(shù)據(jù)，并且，在早期發(fā)展的進程中給任何不合需要的加亮燈設計特點。雖然，商業(yè)的CFD包裹以前習慣依靠震動壓力來預測時間，但是，計算設備似乎已經(jīng)限制于大量工作而僅模擬渦螺和葉輪的交互作用。考慮到?jīng)]有吸入口的輸入支管和泄流流程。當前的工作目的是去改進和以前的工作范圍，涉及由于模擬產(chǎn)生的震動壓力，包括完整的水泵幾何學。</p><p>　　數(shù)字模擬合

42、并了所有泵中主修的流動流程，環(huán)繞著吸入口，葉輪，泄露管徑和渦螺保護套。工作集中于縮小在雙力高壓葉輪的比例，單向多極泵的排列。通過泵的制造商對泵震動壓力的理解似乎驚人的有限。對于泵中震動壓力安全水平?jīng)]有正式的標準，只有工業(yè)上會采用保證比泵出口壓力小百分之3的方針。然而，不知道一個在泄流處相應百分之3的限度在其他泵內(nèi)裝置上是否是安全限度。因此，在葉輪和渦螺中對震動壓力詳細的估計已經(jīng)完全被測量，無論在泄流處百分之三的限度是否被采用，實際上都已

43、給與泵的其他主要機件所有保證。分析包括三個流動速度超過流動范圍從一個要求泵輸送量流速范圍伸出到一個普通泵最小操作點的百分之25。這里的分析只是大量參數(shù)的一部份，研究泵內(nèi)部震動壓力幾何學的影響。這次工作的另一個目的是表示震動壓力合理的估計和震動壓力能夠被獲得的傾向，因為不同的泵的幾何學，通過數(shù)字表示的方法在一個合理的期限內(nèi)。</p><p><b>　　2、試驗研究</b></p>

44、<p>　　工業(yè)試驗者可以利用一些試驗的結(jié)果，檢查縮小比例泵的震動壓力。這些試驗在數(shù)字分析之前就使用了很多年了。泵被承認的流動是從一個閉式系統(tǒng)，這種泵位于上游3.5個距離，下游4個距離。直接的在上游彎曲之后，直流被使用于為了減少流動因素導致流入泵中。十個Z類型的壓力是裝在泵上的。</p><p>　　壓力被用來收集數(shù)據(jù)，在各個固定地點周圍的泵?？足@在特定的地點靠近泵和管用于連接壓力傳感器，以每個地點

45、。路徑距離出鐵點指向該傳感器是保持盡可能短（ 10-15毫米） ，以確保任何共振頻率所造成的路徑距離是上述的測量范圍。</p><p>　　一Z型壓力傳感器是利用在實驗測試。這個傳感器是安裝在葉輪遮掩物15毫米背后的吸力面對一個葉輪葉片。電氣信號從傳感器被移送從旋轉(zhuǎn)元素，固定數(shù)據(jù)記錄器通過滑環(huán)安排在非驅(qū)動器末端泵。該試驗臺是運行在一個初步的速度，每分鐘轉(zhuǎn)速超過1400年的流量范圍25-125 ％，與5分鐘錄音的傳

46、感器正在采取在每25 ％的流量增量。該1.00 條件等同流速500立方米每小時和象征性帶頭點33.8米為內(nèi)插時，葉輪泵的運行在1400 RPM的。流量變化所取得的調(diào)整一流量閥，使它盡可能的緩慢、平滑。一系列的測試，表現(xiàn)為一些泵的幾何安排，但只有兩個是在這里特別有興趣的（被稱為A和B作為詳細以上） 。一個時間歷史的壓力變化，在上述每個地點都是有記錄的。 1譜分析，當時進行的這一數(shù)據(jù)與壓力脈動被輸出作為的均方根（ RMS的）壓力脈動。 由

47、于時間限制，所涉及的實驗測試，詳細的性能數(shù)據(jù)沒有考慮。測試工作的進行，在按照正常的工業(yè)做法，這是認為，壓力是準確的± 1.5 ％ 。</p><p>　　3、數(shù)值調(diào)查利用CFX終端區(qū)域時序控制流量是一個商業(yè)三維的CFD代碼，采用有限元為基礎的有限體積法，以運輸方程。這種方法提供，保留幾何的靈活性，有限元方法，同時保留養(yǎng)護性能的影響有限體積法，即低的數(shù)值誤差對非光滑的網(wǎng)格。該守則擁有在同一地點的結(jié)構(gòu)網(wǎng)格

48、和耦合的代數(shù)多重計劃，以解決系統(tǒng)的方程。這是一個完全隱式求解方法，因此，造成沒有時間和步驟的限制，被認為很容易執(zhí)行。這并不是任何影響穩(wěn)定的國家的解決辦法，但這限制暫態(tài)計算，只有第一階準確的時間。該利用CFX終端區(qū)域時序控制流量求解也是一個耦合求解，即勢頭和連續(xù)性方程的解決同時進行。這種做法減少了迭代次數(shù)須獲得的收斂性和沒有壓力校正來說，是需要保留的質(zhì)量守恒，導致了更強有力的和準確的求解。該計劃還包括一些會前和會后處理能力，是專門面向葉輪

49、機械零件;這些方便成立了該模型和考試的結(jié)果。利用CFX終端區(qū)域時序控制流量已經(jīng)經(jīng)過驗證的追蹤記錄在葉輪機械的應用，項目眾多的文學出版。</p><p>　　3．1、模型生成葉輪幾何創(chuàng)建使用利用CFX刀刃作為兩個鏡像邊，使用最大直徑366毫米。牛鼻方面的內(nèi)插葉輪設計證明，很難捕捉與利用CFX刀刃，作為該計劃并非設計用來模型雙進入葉輪，并且不允許為樞紐的配置文件，終止在牛市的鼻子安排。為了避免這個問題，雖然真正的樞

50、紐，有效地完成在點對稱性，一個虛構(gòu)的樞紐延長至中點牛鼻葉輪出口沿葉輪線的對稱性。因此，只有輕微的修改，經(jīng)向樞紐簡介一個令人滿意的模型制作內(nèi)嵌葉輪。交錯葉輪可用于無任何修改，由于子午流徑，任何一方的葉輪不連接。</p><p>　　利用CFX網(wǎng)格來生成網(wǎng)格周圍的葉片對任何一方的葉輪，與網(wǎng)格兩邊的葉輪被完全相同。為內(nèi)插葉輪，護理是采取措施，確保網(wǎng)格是在徑向的立場牛鼻。這是必需的援助，建立一個界面之間的葉輪在利用CFX

51、終端區(qū)域時序控制流量，但它也限制了高質(zhì)量的網(wǎng)格。與初步葉輪網(wǎng)格估計約五十萬節(jié)點，這是立即明顯的分裂，這12個網(wǎng)格之間的葉輪通道和泄漏流將導致在葉輪通道網(wǎng)格低于理想的大小。一些已作出努力集中于較小的網(wǎng)格大小，而包括網(wǎng)格大如86499節(jié)點。由于葉輪的互動與蝸殼是非常重要的，決定進行網(wǎng)格的獨立檢查是否使用了穩(wěn)定狀態(tài)的分析納入模型構(gòu)成的一半，一個水泵，即6葉輪通道和二分之一泵蝸殼（利用對稱性）。蝸殼模型是一致的，在每個網(wǎng)格的獨立分析，與分布相似

52、，所用的最后泵模型?？梢灶A料，這種相互作用會引起較大的分歧很明顯之間的網(wǎng)格比通常顯示在網(wǎng)格獨立的葉輪，只有比較，由于日趨復雜的流通模式。由分析來說，在工作地點的水流條件（1.00qn），邊界條件，正在大規(guī)模流進和平均靜態(tài)壓力的插座，與最高殘余收斂準則被設定為1e-4（最大值） 。壓力數(shù)據(jù)報告，下面涉及到的變異在一個單一的葉輪通道（從中間通過）在一個單一的地位，蝸殼。</p><p>　　3.2、解決參數(shù)由于該議

53、案的葉片相對平穩(wěn)蝸殼是中環(huán)至調(diào)查，分析必須涉及多個參照系。 ，以防止兩個轉(zhuǎn)子/定子接口正在互相垂直于葉輪出口，泄漏流的路徑和葉輪的網(wǎng)格定在轉(zhuǎn)動的參照系。吸力進氣道和蝸殼網(wǎng)格在靜止的參照系。網(wǎng)格接口中使用的研究可歸納如下： 內(nèi)部組件方面：一般的網(wǎng)格界面（ ggi ） 兩國固定組成部分：一般的網(wǎng)格界面之間的兩個轉(zhuǎn)動分量：一般的網(wǎng)格界面之間的旋轉(zhuǎn)和固定部分組成：  凍結(jié)轉(zhuǎn)子接口（穩(wěn)態(tài)分析）  轉(zhuǎn)子/定子接口（瞬

54、態(tài)分析） 被凍結(jié)的轉(zhuǎn)子界面實現(xiàn)框架沒有改變的相對位置隨時間變化和界面平均。本地流動的特點是允許運輸全國界面，因此，壓力非一致結(jié)構(gòu)允許宣傳。身體的比喻是想象觀測流量之間的固定和旋轉(zhuǎn)構(gòu)件下的頻閃照明。轉(zhuǎn)子/定子界面是一個滑動（幀改變）接口，可以帳目瞬態(tài)的互動效果。網(wǎng)格兩側(cè)的轉(zhuǎn)子/定子界面始終是在相對運動與尊重，到另一個。 有三種常見的組合邊界條件往往是用于分析水泵流量。 進氣道：總壓出路：質(zhì)量流量進氣道：質(zhì)量流量出路：靜壓力進氣道