外文翻譯--熱連軋機(jī)震顫的識別和解決對策

1、<p>　　熱連軋機(jī)震顫的識別和解決對策</p><p><b>　　1導(dǎo)論</b></p><p>　　對任何軋機(jī)來說保證質(zhì)量和生產(chǎn)力是主要的要求。無論是操作還是維修人員，目標(biāo)都是保證成品的性能和外觀。實(shí)現(xiàn)這些目標(biāo)必須要求所有的設(shè)備都正常運(yùn)行。</p><p>　　在熱、冷連軋工藝中，機(jī)械振動在大多數(shù)情況下是可見的，某些特定的條件下

2、甚至產(chǎn)生震顫。由于這個震顫,軋機(jī)機(jī)座的出口處橫向標(biāo)志覆蓋寬度都留下了明顯的痕跡,。過去研究人員已經(jīng)注意到在冷軋線發(fā)生這種現(xiàn)象,但問題并不像熱軋中那么廣泛。這種情況是極其有害的,因?yàn)樗粌H改變了產(chǎn)品的表面質(zhì)量,而且在某些嚴(yán)重的情況下使儀表也發(fā)生了震顫</p><p>　　Siderar的66英寸熱連軋生產(chǎn)線包含四個加熱爐、一個粗軋機(jī)、除鱗機(jī)，生產(chǎn)速度每分鐘640公尺,產(chǎn)品范圍從16~1250mm厚、291~560m

3、m寬。一個熱軋廠的概圖顯示。在處理小斷面小寬度的馬口鐵產(chǎn)品時，重復(fù)發(fā)生振動問題，并由此,工作輥磨損增加,影響效率，次品率也在增加。這個問題是只發(fā)生在第二架軋機(jī)處，軋輥和軋制速度每分鐘接近120米時。一般來說,這個振動伴隨著一個強(qiáng)度變化的低頻的嗡嗡聲。</p><p>　　1997年8月,Siderar決定建立一個與尺寸數(shù)據(jù)相關(guān)的計(jì)算模型來尋找震顫的源頭。而從以前的工作,主要是一系列的測量軋制速度和觀測主軸齒輪輪齒

4、的振動特性。</p><p>　　為了改進(jìn)對發(fā)生在Siderar的軋機(jī)上的現(xiàn)象的理解，我們做了一個額外的測量和模式分析研究。在二月,進(jìn)行第二次測量來大體上完成一個顫動的波譜特性描述,包括運(yùn)動系統(tǒng)和軋制法測定了裝配的固有頻率和驗(yàn)證模型,發(fā)現(xiàn)在各個不同點(diǎn)中存在相關(guān)性。在后者情況下,同時計(jì)算了驅(qū)動系統(tǒng)中的扭轉(zhuǎn)頻率和立輥的振動的簡化模型,模擬了大型有限元分析模型。</p><p>　　在2000年

5、5月, Anther測量了軋制馬口鐵軋件時，第二和第三軋機(jī)處的相關(guān)數(shù)據(jù)。</p><p>　　從計(jì)算結(jié)果與實(shí)測值的比較,給出了一個對這些現(xiàn)象的解釋。</p><p><b>　　2所描述的問題</b></p><p>　　在Siderar軋機(jī)最終軋制中的顫振問題可追溯到軋機(jī)的啟動,并經(jīng)歷了相當(dāng)數(shù)量的增加，因?yàn)閮蓚€因素:</p>&

6、lt;p>　　(1)介紹了在F2和F3軋機(jī)外殼的改良，即為變輥和彎輥的移動塊增加了窗口。</p><p>　　(2)對產(chǎn)品表面質(zhì)量的要求增加，放任了震顫缺陷。在軋制輕薄材料時最大的振動條件出現(xiàn)了,比如馬口鐵, 或在遲于材料滾軋時間表時,如1020mm寬、1.60mm厚的。</p><p>　　產(chǎn)生這一現(xiàn)象可能突然的,在某些情況下是在軋制的第一卷時,在其他情況中,也可能是在軋制過程中經(jīng)

7、過逐漸變化后達(dá)到最大幅度。當(dāng)振顫在首軋時產(chǎn)生后，隨后進(jìn)入第四，第五和第六軋機(jī)會修正表面差異。在最不利的情況下，基于觀察到成品樣品上的問題會需要一次換輥。</p><p>　　被移走的軋輥會通過不同強(qiáng)度和數(shù)量的振動頻率顯示在整個寬度表面上的振顫。類似的，在第二和第三軋機(jī)的出口處帶卷的上表面，就能觀察到振顫的標(biāo)志。但是大多數(shù)軋輥的不均勻磨損發(fā)生在下輥。</p><p><b>　　3

8、進(jìn)行測量</b></p><p>　　進(jìn)行了三套測量: 1997年8月,1998年2月,1998年7月。在第二架軋機(jī)處，最初的兩個表示振動行為的特征通過加速表的信號頻譜分析在圖3中顯示。</p><p>　　在第一個數(shù)據(jù)集，僅僅鎖定了每個信號的線性譜,目的就是確定主軸的磨損是否是振顫最主要的起因。在這種情況下,通過固有頻率的計(jì)算分析了基于關(guān)聯(lián)測量光譜特征頻率的運(yùn)動學(xué)系統(tǒng)。得到如

9、圖所示的一個特征譜圖。</p><p>　　主軸的牙齒嚙合運(yùn)行過程中,大約60r/min轉(zhuǎn)速時，觀察到一個頻率大約40到41赫茲的頻率。當(dāng)軋制厚板帶時,第二架軋機(jī)的速度大約降低25%和沒有產(chǎn)生振顫。因此,我們可以推斷一個來源于磨損的主軸的輪齒的周期的力激起了共振。下輥比上輥的振動更強(qiáng)烈是由于下輥剛度比上輥大。此外,高的主軸剛度更有容易將振動傳遞到工作輥。作用第二架軋機(jī)運(yùn)動系統(tǒng)部分的主動力和第一組數(shù)據(jù)中產(chǎn)生過程變量

10、可以見表。</p><p>　　為了驗(yàn)證結(jié)果,進(jìn)行了另一種更完整的測量,包括對不同的振動的波形進(jìn)行了測定以及光譜分析與對比。通過加速度計(jì)和鄰近的一個放在上工作輥和移動塊之間的傳感器測量，經(jīng)過惠普3560A分析儀得到數(shù)據(jù)。</p><p>　　1998年7月進(jìn)行的測量目的就是利用技術(shù)來確定第二架軋機(jī)的固有頻率。</p><p>　　估算中的傳遞函數(shù)在每一個測試圖中的分

11、析真實(shí)的和虛構(gòu)的自然的頻率的方法，顯示出各方面垂直測量的函數(shù)的大小。頻率為60.75 和 72.75 赫茲時共振是最明顯的。在圖中，大量的傳遞函數(shù)表明，在垂直方向存在一個橫向激勵。在這種情況下,主共振在10至90赫茲的范圍內(nèi)大約在39.5赫茲時很明顯。</p><p>　　在2000年5月,另一個是在第二第三架軋機(jī)的正常的工藝驗(yàn)證條件,利用相同的測試點(diǎn)評估表格中主動力的影響。結(jié)果表明了實(shí)質(zhì)性的進(jìn)步。</p&

12、gt;<p><b>　　4計(jì)算模型的結(jié)果</b></p><p>　　說明下測量方法，通過不同的形式執(zhí)行一個關(guān)于第二架軋機(jī)的特性模型。主要的目的是為了確定運(yùn)動學(xué)系統(tǒng)的哪些部分在振顫的共振中表明一種共振條件,如光譜特性。再者，在軋制帶材時發(fā)現(xiàn)了振顫痕跡，這只有軋輥和機(jī)架的垂直移動能產(chǎn)生，正如兩者覆蓋了帶卷寬度。主要的前提是尋找一種形式的在主軸上放大的扭轉(zhuǎn)振動,源于它是熱軋中的振

13、動的主要原因。最復(fù)雜的計(jì)算是確定正常頻率和振型的正常利用振型的有限元法的整體結(jié)構(gòu)。運(yùn)用整個體系結(jié)構(gòu)分析程序,區(qū)分機(jī)架的主要部分為有限的立體元件的類型、模擬六面體旋轉(zhuǎn)構(gòu)件與梁式的元素。</p><p>　　該模型包括了主軸、上分離器和鋼柱支承。表格顯示最重要的計(jì)算的頻率。例如,第三振型在35赫茲。</p><p>　　為了估計(jì)傳動體系中的扭轉(zhuǎn)頻率,從工作輥和支承輥經(jīng)過整個運(yùn)動系統(tǒng)直到電動機(jī),

14、為整個傳動體系建模，劃分匯編成大約1100元素,最大的數(shù)目為主軸。用Holzer法來解決問題。頻率的影響顯示在表格中。進(jìn)行了靈敏度研究,改變組件和慣性矩。計(jì)算結(jié)果影響不大,這表明改變固有頻率也有必要做出重大修改裝配、堆疊裝配的四個軋輥的垂直振動頻率。這是基于一項(xiàng)類似于利用參考的方案。主要結(jié)果顯示在表格。</p><p><b>　　5解釋測量結(jié)果</b></p><p&g

15、t;　　現(xiàn)場檢測的非臨界的資料顯示軋制過程中主要的頻率,13、17、32、55Hz沒有哪個激起機(jī)架的特征頻率。</p><p>　　在窄帶的軋制中，出現(xiàn)在主頻35.7赫茲。另一方面, 在固有頻率測量中經(jīng)由影響技術(shù),這個影響是:39.5和60.5赫茲,代表機(jī)架的垂直運(yùn)動；84.5赫茲，相當(dāng)于軋輥的垂直運(yùn)動；大約41赫茲,由于工作輥和齒輪耦合的切向沖擊,用軸向的傳感器測得。</p><p>　

16、　振動就會發(fā)生在接觸弧的摩擦條件變得不穩(wěn)定（即速度增加摩擦系數(shù)減?。?。當(dāng)擾亂發(fā)生,扭矩增加,產(chǎn)生了一種自我維持的振動。這個結(jié)果可以聯(lián)系到大幅衰減和摩擦滑脫的條件。</p><p>　　扭轉(zhuǎn)頻率焦點(diǎn)一般落在40到41赫茲，而機(jī)架的垂直振動在35赫茲。這些價值就是, 考慮到基于Siderar關(guān)于振顫特性的假設(shè)的簡化描述，測量軋制馬口鐵材料中發(fā)生的振顫能得到頻率為37.5赫茲。</p><p>

17、　　振動在板帶進(jìn)入軋機(jī)時產(chǎn)生，可能是由潤滑不足(摩擦系數(shù)) 產(chǎn)生的軸扭轉(zhuǎn)振動。這些主軸扭轉(zhuǎn)振動在機(jī)架上激起一個垂直振動模式,就像驅(qū)動了一個運(yùn)動放大器，通過計(jì)算有限元模型可能相當(dāng)于表格中的振動模式三。這個運(yùn)動是從軋輥開始的。一旦這種情況發(fā)生時,由于振顫的工作輥中的振動，其效果是持續(xù)的。</p><p><b>　　6 結(jié)論</b></p><p>　　通過完美地識別源頭

18、來成功地止住熱軋開始時的振動。這些測試進(jìn)行分析,使Siderar得到的結(jié)果。</p><p>　　振顫是一個出現(xiàn)在操作中的軋機(jī)上的特異的能自我維持的振動,由于三者之間的動態(tài)結(jié)構(gòu)的設(shè)備裝配和軋制過程本身。</p><p>　　敏感的分析顯示實(shí)際上不可能在這個時刻修改大部分被牽連部分的慣性。</p><p>　　最有效的對策是,降低潤滑油的間隙的移動組件和改進(jìn)的頻率變化

19、的主軸。結(jié)果，這些措施大大降低熱軋振動。</p><p>　　當(dāng)主軸在這個狀態(tài)的時候，想關(guān)心減少，溫度和板帶速度下降。</p><p>　　Identification and Countermeasures to Resolve Hot Strip Mill Chatter</p><p>　　Introduction</p><p>　　

20、Quality and productivity are major requirements for any rolling mill. Physical and metallurgical properties and surface appearance of the finished product are goals shared both by the operating and maintenance personnel.

21、 All equipment must function properly to achieve these goals.</p><p>　　In both hot and cold rolling processes, mechanical vibrations called chatter can be observed under certain condition, in most cases audi

22、bly. As a result of this chatter, transverse marks covering the strip width are impressed at the exit of the considered mill stand. The occurrence of this phenomenon in the last stands of cold rolling mills has been repo

23、rted by several researchers, but information is not as widespread foe hot mills. The problem is highly undesirable, as it not only changes strip</p><p>　　Siderar’s 66-inch hot strip mill consist of four rehe

24、ating furnaces, a condition roughing mill, a descaling mum speed of 640 meter per minute, Product ranges from 1.6 to 12.5 mm in thickness and 560 to 1525 mm in width. A schematic of the hot strip mill facility is shown i

25、n Fig. During the processing of light gauge and narrow tinplate product, a repetitive vibration problem occurred. Consequently, work roll wear increased, affecting mill productivity. Reject rates also increased. This pro

26、blem wa</p><p>　　In Aug. 1997, Siderar decided to carry out an intensive campaign of measurements and data correlation with calculation models to detect chatter origin. From previous work, a series of measur

27、ements was made to relate rolling speed and number of spindle teeth with observed vibration characteristics.</p><p>　　To improve understanding of the phenomenon at Siderar’s rolling mill, an additional set o

28、f measurements and a mode analysis study were performed. On Feb a second set of measurements was carried out for achieving a spectral characterization of the vibrations in the whole stand, as well as the existing correla

29、tion among different points, including an analysis of the kinematic system and rolling for determination of the assembly’s natural frequencies and validation of calculation models. In the lat</p><p>　　In May

30、 2000, Anther set of measurements was carried out on stands F2 and F3, during rolling of tinplate bands.</p><p>　　Results obtained from calculations and their comparison with measured values, as well as an e

31、xplanation of the observed phenomenon observed on these results, are presented.</p><p>　　Description of the problem</p><p>　　The chatter problem in the finishing mill at Siderar dates back to th

32、e mill’s start-up, and experienced a sizeable increase due to two factors:</p><p>　　Modifications were introduced in the F2 and F3 stands-housing windows were machined for installation of mobile blocks for r

33、oll shifting and roll bending.</p><p>　　Increased demand for surface critical product, free of chatter defects. The conditions of maximum vibration appeared when light gauge and thin material was being rolle

34、d, material for tinplate, or in material rolled late in the rolling schedule, such as 1020 mm wide and 1.60 mm thick.</p><p>　　Appearance of the phenomenon could be sudden, in some cases during the first str

35、ips of the rolling schedule, and in other case after a gradual evolution, reaching its maximum level in the middle of the rolling. As the chatter affected the fist stands, subsequent passes in the F4, F5 and F6 stands co

36、rrected surface variations. In the most unfavorable cases, a partial roll change was required based on observed wear on finished product samples.</p><p>　　The removed rolls showed chatter on the whole surfac

37、e width, with varying intensity and number of marks related to the vibration frequency. Similar to others, chatter marks were observed on the upper surface of the strip at exit of stands F2 and F3, but the roll most affe

38、cted by irregular wear was he lower roll.</p><p>　　Measurements performed</p><p>　　Three sets of measurements were performed: Aug 1997, Feb. 1998 and July 1998. The first two characterized the v

39、ibratory behavior of the F2 stand via spectral analysis of signals measured with accelerometers in the locations indicate in Fig.3.</p><p>　　In the first data set only the linear spectra of each signal were

40、determined, as the intention was determine if the primary cause of the chatter was spindle wear. In this case, analysis of the measured spectra was based on correlation of the characteristic frequencies of the kinematic

41、system with the natural frequencies calculated. A characteristic spectrum is illustrated in Fig.</p><p>　　One observed frequency was approximately 40 to 41 Hz, which corresponds to the spindle teeth meshing

42、during operation around 60 rpm. When thicker strips were rolled, the F2 speed is around 25% lower and no chatter was observed. Thus, it can be concluded that resonance was excited by a periodic force originating in worn

43、spindle teeth. A higher vibration in the lower roll than the upper roll was due to a higher stiffness of the lower part of the housing compared with the upper part. </p><p>　　Additionally, the high stiffness

44、 of the spindle more efficiently transmitted vibrations to the work roll. </p><p>　　Actions taken on elements of the F2 kinematic system and process variables resulting from the first set of measurements are

45、 presented in Table.</p><p>　　TO validate the results, a second and more complete set of measurements was performed, including wave shape determination, spectral analysis and correlation of vibration from at

46、 different measurement points. </p><p>　　Measurement were obtained with accelerometers and proximity sensors positioned between the upper work roll and the mobile shifting blocks, together with a Hewlett Pac

47、kard 3560A two-channel analyzer.</p><p>　　The measurements performed in July 1998 had the purpose of determining the natural frequencies of the F2 stand utilizing the impact technique.</p><p>　　

48、Methodology for determining natural frequencies consisted of analyzing the real and imaginary parts of the estimated transfer function in each of the tests Fig shows the magnitude of the function for a vertical measureme

49、nt on each stand. Resonance frequencies of 60.75 and 72.75 Hz are evident. In Fig, the magnitude of the transfer function, take in the vertical direction in presence of a horizontal excitation, is shown. In this case, th

50、e main resonances were within the range of 10 to 90 Hz and t</p><p>　　In May 2000, another verification was performed under normal process conditions on stands F2 and F3, utilizing the same measurement point

51、s to evaluate the influence of the action items taken in Table. Results indicated a substantial improvement.</p><p>　　Results of the calculation model</p><p>　　To interpret the measurements, a m

52、odal characterization of stand F2 was carried out with different models. The main objective was to determine which part of the kinematic system enters into resonance during chatter, as spectra characteristics indicated a

53、 resonance condition. Furthermore, chatter marks observed on the rolled strip indicated that only vertical movements of the roll/stand assembly could produce them, as they covered the entire strip width. The predominant

54、hypothesis was to search f</p><p>　　The model includes the spindle, upper separators and mass of the cylinder supports. Table shows the most important calculated frequencies. For example, the third mode of v

55、ibration at 35 Hz.</p><p>　　For estimation of the torsional frequencies of the drive assembly, from the work and backup roll through the kinematic system up to the motor, the drive assembly was modeled, divi

56、ding the assembly into approximately 1100 elements, the greatest number being for the spindle. The Holzer method was used to solve the problem. The frequencies of interest are indicated in Table. A sensitivity study was

57、performed, changing component masses and moments of inertia. Little influence on calculation results </p><p>　　5 Interpretation of the measurement results</p><p>　　Field measurements taken durin

58、g rolling of noncritical material revealed predominant frequencies of 13, 17, 32 and 55Hz, none of which excites any of the characteristic frequencies of the stand.</p><p>　　During narrow strip rolling, a pr

59、edominant frequency of 35.7Hz occurs. On the other hand, among the natural frequencies measured via the impact technique, those of interest are: 39.5 and 60.5Hz, representing a vertical movement of the stand; 84.5Hz,corr

60、espond to the vertical movement of rolls; and approximately 41Hz, resulting from tangential impact on the coupling between the work roll and the pinion stand, using a sensor tangential to the shaft.</p><p>　

61、　Vibrations result when friction condition in the arc of contact are unstable (i.e, where the friction coefficient decreases as speed increases). When a disturbance occurs, torque is increased torque, producing a self-su

62、stained vibration. This result relates to high reductions and conditions of friction with slip.</p><p>　　Focus generally falls on torsional frequencies of 40 to 41 Hz and vertical vibration of the stand at 3

63、5 Hz. These values, when associated with the 37.5 Hz frequency measured during chatter when rolling material for tinplate, allow for a simplified description based on the following hypothesis for chatter characteristics

64、at Siderar.</p><p>　　Vibration is incited when strip enters the stand and is probably roduced by inadequate lubrication (friction coefficient), yielding torsional vibrations of the spindle. These torsional s

65、pindle vibrations excite a vertical vibration mode of the stand, which actuates as a movement amplifier, likely corresponding to vibration mode 3 in Table calculated by the finite element model. This movement starts maki

66、ng the roll. Once this occurs, the effect is perpetuated by vibration of chattered work roll.</p><p>　　Couclusions</p><p>　　Successfully suppressing vibration in hot rolling starts with perfect

67、identification of the source. Analytical tolls and field measurements performed allowed Siderar to converge on the obtained result.</p><p>　　Chatter is a peculiar self-sustained oscillatory movement that app

68、ears during the operation of rolling mills, due to interaction between the dynamic structure of the equipment assembly and the rolling process itself.</p><p>　　Sensitivity analyses revealed the practical imp

69、ossibility of modifying the moments of inertia of the most compromised element.</p><p>　　The most efficient countermeasures were lubricated rolling, decrease of the clearance of the mobile assemblies and mod

70、ification to the frequency of spindle changes. As a result of these corrective action, hot strip mill vibration decreased considerably.</p><p>　　When spindles are in phase, a correlation among reduction, spe