ANSUHI 9.6.4-2000 ANSUHE 9.6.4-2000 @ American National Standard for Centrifugal and Vertical Pumps for Vibration Measurements and Allowable Values HydraulicANSI/HI 9.6.4-2000 American National Standard for Centrifugal and Vertical Pumps for Vibration Measurements and Allowable Values Secretariat Hydraulic Institute www pumps.org Approved March 23, 2000 ‘American National Standards Institute, Inc. oe HydrauieInstute tancards, Copyright © 1097-2005, At Rights Reserved7 American 4er0val of an American National Standard requires verification by ANSI that the National Standard Publishes By requirements for due process, consensus and other criteria for approval have been met, by the standards developer. Consensus is established when, in the judgement of the ANSI Board of Standards Review, substantial agreement has been reached by directly and materially affected interests. Substantial agreement means much more than a simple majority, ut not nec- essary unanimity. Consensus requires that all views and objections be considered, and that a concerted effort be made toward their resolution. The use of American National Standards is completely voluntary; their existence does not in any respect preclude anyone, whether he has approved the standards or not, from manufacturing, marketing, purchasing, or using products, processes, or proce dures not conforming to the standards. ‘The American National Standards Institute does not develop standards and will inno circumstances give an interpretation of any American National Standard. Moreover, no person shall have the right or authority to issue an interpretation of an American National Standard in the name of the American National Standards Institute, Requests for interpretations should be addressed to the secretariat or sponsor whose name ‘appears on the ttle page of this standard. CAUTION NOTICE: This American National Standard may be revised or withdrawn at any time. The procedures of the American National Standards Institute require that action be taken periodically to reatfirm, revise, or withdraw this standard, Purchasers of ‘American National Standards may receive current information on all standards by call- ing oF writing the American National Standards Institute Hydraulic institute 9 Sylvan Way, Parsippany, NJ 07054-3802 www.pumps.org Copyright © 2000 Hydraulic Institute All rights reserved. "No part ofthis publication may be reproduced in any form, in an elecyonic retrieval system or otherwise, without prior \ntton permission ofthe publisher. Printed in the United States of America ISBN 1-880952-45.9 ysrou institte Standards, Copyiht © 1997-2005, All Rights ReservedContents. fl a Foreword v . 9.6.4.2.6 Structure dynamic analysis (optional) T = 9.6.4.3.2 Location of vibration probes 8 a ‘Appendix B Index 26 Fures 2.844 Ent scion et moved ANS! B72 4,873.3, 8795 8 9.6.4.1 —End suction, hard metal and rubber-lined horizontal and vertical .. 16 916.4.12 — Between bearing, single and multistage 17 Hydraui Inetute Standards, Cpyight © 1997-2006, Al Rights Reserved iBEEERE EERE EBE EEE EERE EERE REE EEE 9.6.4.13 — Vertical turbine, mixed flow and propeller type 9.64.14 — Vertical turbine, short set pumps, assembled for shipment by the manufacturer 19 9.6.4.15A — Allowable residual unbalance in pump impellers grade G6.3 (metric) 21 9.6.4.15B — Allowable residual unbalance in pump impellers — grade 66.3 (US units) 22 iv yore Intute Standards, Copyright © 1997-2005, Al Rights ReservedForeword (Not part of Standard) Purpose and aims of the Hydraulic institute The purpose and aims of the Institute are to promote the continued growth and well-being of pump manufacturers and further the interests of the public in such matters as are involved in manufacturing, engineering, distribution, safety, trans- portation and other problems of the industry, and to this end, among other things: 1). To develop and publish standards for pumps: ) To collect and disseminate information of value to its members and to the public; €) To appear for its members before governmental departments and agencies ‘and other bodies in regard to matters affecting the industry, ) To increase the amount and to improve the quality of pump service to the public ) To support educational and research activities: ) To promote the business interests of its members but not to engage in bust ress ofthe kind ordinarily carried on for profit or to perform particular services for its members or individual persons as distinguished from activities 10 improve the business conditions and lawful interests of all of its members Purpose of Standards 1) Hydraulic Institute Standards are adopted in the public interest and are designed to help eliminate misunderstandings between the manufacturer, the purchaser and/or the user and to assist the purchaser in selecting and obtaining the proper product for a particular need, 2). Use of Hydraulic Institute Standards is completely voluntary. Existence of Hydraulic Institute Standards does not in any respect preclude a member from manufacturing or selling products not conforming to the Standards. Definition of a Standard of the Hydraulic Institute Quoting from Article XV, Standards, of the By-Laws of the Institute, Section B: ‘An Institute Standard defines the product, material, process or procedure with reference to one or more of the following: nomenclature, composition, construc tion, dimensions, tolerances, safety, operating characteristics, performance, qual- ity, rating, testing and service for which designed. Comments from users Comments from users of this Standard will be appreciated, to help the Hydraulic Institute prepare even more useful future editions. Questions arising from the con- tent of this Standard may be directed to the Hydraulic Institute. It will direct all ‘such questions to the appropriate technical committee for provision of a suitable answer. Ita dispute arises regarding contents of an Institute publication or an answer pro- vided by the Institute to a question such as indicated above, the point in question shall be referred to the Executive Committee of the Hydraulic Institute, which then shall act as a Board of Appeals Hydra Inetite Standards, Copyright © 1967-2005, All Rights Reservad vRevisions The Standards of the Hydraulic Institute are subject to constant review, and revi- sions are undertaken whenever itis found necessary because of new develop- ments and progress in the art. Units of Measurement This standard is written using both metric and US Customary units of measure- ment. Metric units appear frst followed by US units in brackets, Tables, charts and sample calculations are printed twice, first in metric units, then in US units. Consensus for this standard was achieved by use of the Canvass Method The following organizations, recognized as having an interest in the standardiza- tion of centrifugal pumps, were contacted prior to the approval of this revision of the standard. Inclusion in ths ist does not necessarily imply that the organization concurred with the submittal of the proposed standard to ANSI. AR. Willey & Sons, Inc ‘Afton Pumps, Inc. ‘ANSIMAG Incorporated Bechtel Corporation Black & Veatch LLP Brown & Caldwell Carver Pump Company Cascade Pump Co has. S. Lewis & Company, Inc. Chempump Division, Crane Pumps & Systems Cheng Fluid Systems, inc Cuma S.A Dean Pump Division, Metpro Corp. DeWante & Stowell Dow Chemical Essco Pumps Exeter Energy Limited Partnership Fairbanks Morse Pump Corp. Fortis State Univ. Construction and Facilities Dept Flowserve Corporation Fluid Sealing Association Franklin Electric Grundfos Pumps Corporation Ilinois Department of Transportation ITT Fluid Handling (B & G) ITT Fluid Technology ITT Flygt Corporation Iwaki Walchem Corporation J.P. Messina Pump and Hydr. Cons. John Crane, Inc. Krebs Consulting Service KSB, Inc. Lawrence Pumps, Inc. MW. Kellogg Company Malcolm Pirnie, Inc. Marine Machinery Association Marshall Eng. Prod. Co. (MEPCO) Moving Water Industries (MW) Ortev Enterprises Inc. Pacer Pumps Pacheco Engineering Patterson Pump Company Pinellas County, Gen. Serv. Dept. Price Pump Company Raytheon Engineers & Constructors Red Jacket Reddy-Butfaloes Pump, Ine. Scott Process Equipment Corp. Seller Supply Company ‘Skidmore South Florida Water Mgmt. Dist. Sta-Rite Industries, Inc Sterling Fluid Systems (Canada) Inc. Stone & Webster Eng. Corp. ‘Summers Engineering, Inc Systecon, Inc. Taco, Inc The Process Group, LLC University of Montana Val-Matic Valve & Manufacturing Corp. Yeomans Chicago Corporation Zoeller Engineered Products “ Hydraulic nsttute Sardards. Copyight © 1997-2005, Al Rights ReservedAlthough this standard was processed and approved for submittal to ANSI by the Canvass Method, a working committee met many times to facilitate the develop- ment of this standard. At the time it was developed, the committee had the follow: ing members: Chairman ~ Jack Claxton, Patterson Pump Company Other Members Alternates Thomas Angle, EnviroTech Aleks Roucney, EnviroTech Pumpsystems Pumpsystems William A. Beekman, Floway Pumps Frederic W. Buse, Ingersoll-Dresser Paul Behnke, Ingersoll-Dresser Pump Pump. Michael Derr, Afton Pumps R. Barry Erickson, ITT Industrial Pump Allan Budris, ITT industrial Pump Group Group Herman Greutink, Johnston Pumps John Eddy, Johnston Pumps Gunnar Hovstadius, ITT Flygt Stephan Abelin, ITT Fiyst Al lseppon, Sta-Rite Industries Thomas Morton, Sulzer Bingham Don Spencer, Sulzer Bingham Pat Moyer, ITT, Bell & Gossett Jim Roberts, ITT, Bell & Gossett James Osborne, A.R. Wilfley & Sons Ray Perriman, Sunstrand Fluid Handling Y.J. Reddy, Reddy-Buffaloes Pump ‘Arnold Sdano, Fairbanks Morse Ron Sperry, Flowserve Corporation _—Fred Hery, Flowserve Corporation Rager Turley, Flowserve Corporation Hysraue Instute Standards, Copyright © 1997-2005, Al Rights Ratoned vii9.6.4.1 Introduction/Scope This standard describes the vibration characteristics for industrialcommercial centrifugal and vertical pumps. Included is a description of the dynamics of vibration, vibration measurement, allowable vibration values and factors that effect vibration. 9.6.4.2 Dynamics of vibration Al centrifugal and vertical turbine pumps have rotors ‘and structures that can vibrate in response to excita- lion forces, When the frequency of the excitation forces is close to the natural frequencies of the struc tures, resonance can occur and excessive and dam- aging vibration levels can be reached. These natural frequencies of vibration usually occur in one or more ofthe following modes: + Rotor lateral vibration + Rotor torsional vibration + Structure lateral vibration The natural frequencies of vibration can be deter- ‘mined by one of the following methods: + Simple beam formulas based on those derived from common structural mechanics + Finite elements methods using any one of a num- ber of commercially available computer programs + Experimental techniques using variable frequency exciters or impact devices in conjunction with vibration sensors and recording instruments 9.6.4.2.1 Lateral critical speed ‘The natural frequency of rotor lateral vibration is also called the lateral critical speed. More than one lateral ‘mode can occur naturally, as shawn in the following ‘example of a simple shaft supported by two bearings. In Figure 9.6.4.1, the shat is bowed at the center and Figure 9.6.4.1 — First critical speed HI Pump Vibration — 2000 ibrates back and forth in this shape. in Figure 9.6.4.2, a node or steady point occurs between the bearings, and the shaft takes an 'S" shape and vibrates in this manner, Other more complex shapes may also occur. Figure 9.6.4.1 has the lowest natural frequency and is sometimes called the first critical speed. Figure 9.6.4.2 has a higher natural frequency than Figure 9.6.4.1 and is sometimes called the second cncal speed. Determination of lateral enitical speed is important for pumps and associated rotating equipment because if 2 critical speed or resonant frequency is close to an operating speed or other exciting frequency, such as the impeller vane pass frequency, small excitation forces can be greatly amplified. The resulting stresses. ‘and deflections can cause premature equipment failure Knowledge of critical speed is also important for bal ancing considerations. Rotors having a first critical speed less than the rated rotating speed may require balancing to a more stringent balance level. Calculation of critical speed can become very com- plex, depending on the effects one wants to consider in the calculation. A simple calculation of the frst ertcal speed of a rotor is done by determining the static, deflection of the center of gravity of a shaft or rotor Under its own weight, when assumad to be in a hor zontal position (despite its actual orientation). Knowing the static deflection "dg" in millimeters (inches) of the center of gravity itis possible to calculate the first nats tral frequency or critical speed (in CPM), as follows Metric: N, = 000299 / dy US Units: Ne = 187.7 / dyi® Where: Ne = cfitical speed (rpm) dg = Static deflection, mm (in.) Figure 9.6.4.2 — Second critical speed. Hydra Inetute Standards, Copyeght © 1987-2005, Al Rights ReservedHI Pump Vibration — 2000 If the rotor system is assumed to be simply supported only at its bearings, in air, on infinitely rigid bearing supports, the resulting critical speed is the "dry critical speed.” The following equations yield dry critical speeds for three common configurations. (Note: These ‘examples are intended to be for illustration purposes ‘only and are not intended to provide standard calcula- tion methods.) Examples: Example 1: Shaft, between two bearings dg = (SWL /384E1)° w = shaft weight per length, kNim (pounds per inch) L_ = shaft length, meters (inches) E = modulas of elasticity, GPa (psi) ‘moment of inertia, meters* (in*) RONG oR Bearings Example 2: Impeller between two bearings (massless shaft assumed) [wa? b?/ SEIL]S impeller weight, kN (pounds) E = modulas of elasticity, GPa (psi) 1 = moment of inertia, meters* (in*) L = shaft length, meters (inches) >to ps = ts ih — Example 3: Overhung impeller on two bearings (massless shaft assumed) gy = wa? LIENS W = impeller weight, kN (pounds) E = modulas of elasticity, GPa (psi) J = momentof inertia, meters (in*) L = shafllength, meters (inches) eI Inpeler > Other simple equations are available or can be derived to determine the dry critical speed for the particular case being considered. For many applications, use of dry ortical speed calcu: lation methods are inadequate due to lack of consider- ation of influencing factors. Calculation of "wet critical speed” considers the support and damping effects of the pumped liquid at the operating conditions, bearing flexibility and damping, and other influencing factors. A few of the factors that can effect the critical speed of a Pump and associated equipment that may be consid fered in a wet critical speed analysis include + Stiffness and damping properties of bearings, ‘wear rings, close-clearance bushings, seals and packing + Hydraulic damping effect of liquid surrounding the rotating assembly + Lomakin effect (@ hydrodynamic bearing effect caused by water flowing through small annular areas at wear rings, interstage bushings, balanc: ing devices, etc.) + Weight of the water in the impeller passages + Temperature effects + Wear 2 Hydraulic ineitute Standards, Cpyight © 1997-2005, Al Righls Reserved+ Lubricant properties (ournal-type bearings) + Axial tensioning Gyroscopic effects It simple (dry) calculation methods are used, wider margins of safety are required between operating speeds and calculated critical speeds due to the lack ‘of consideration of these effects (25% Is a typical mar- gin for such methods). Including these effects and determination of higher-order critical speeds are often not practical using simple calculation methods. Appli- cation-specific computer programs are available that calculate multiple critical speeds and consider signif. cant effects in certain cases. Finite element methods used by commercially avail- able computer codes provide flexibility in the ability to analyze a wide range of applications with associated effects and obtain multiple critical speeds. Section 9.84.4 provides a specification by which critical speeds and other natural frequencies may be deter- ‘mined using these methods. Allowable margins of safety between operating speeds and calculated critical speeds vary within the pump industry depending on the service and the complexity of the analysis, Commonly used margins of safety a3 determined by analysis are about 15% to 25%, how- fever, these margins may be impractical in certain applications, such as those having multiple excitation sources and operating at variable speed ‘With respect to difficult applications, margins of safety between operating speeds and actual (not calculated) critical speeds can be less than the 15% to 25% range ‘and allow successful installations to be obtained, depending on the levels of excitation and damping present. The method of calculating critical speed, the extent of the analysis, and the margins of safety to be used should be agreed upon by the purchaser and the manufacturer, Experimental testing to determine critical speed(s) is often possible, although sometimes at considerable expense relative to the value of the equipment. Meth- ods of experimental testing, if required, should be agreed upon by the purchaser and the manufacturer. The party having responsibilty for the pump and driver assembly shall interpret the results of any such tests and make recommendations regarding any required corrective action. HI Pump Vibration — 2000 9.6.4.2.2 Lateral dynamic analysis (optional) The party having responsibilty or the pump and driver assembly may perform a lateral dynamic analysis in the pre-construction stage to determine the natural frequencies and critical speeds of the equipment in the lateral mode and to verify that there are sufficient mar- gins away from the operating speed (or speed range if variable speed), and other excitation frequencies that are characteristic of the equipment. The lateral ‘dynamic analysis model may be performed using finite element analysis (FEA) methods commonly used in commercially avaliable FEA computer codes or other proven computer pragrams, The model shall consider the characteristics of any structural component or mass, and the internal equipment structure as neces- sary to allow the determination of the structure natural frequency characteristics and critical speeds. The analysis shall specifically incorporate any of the fllow= ing items applicable to the system being analyzed: + Pump structure, including internal features such as shaits, bearings, and such features as neces- sary having appropriate strength, mass or damp- ing effects. + Motor or other drive equipment reed frequency. In vertical installations involving drives coupled to the pump, if the motor or other drive equipment reed frequency is not accurately known based on impact test of identical or similar units, a dynamic analysis of the equipment may be performed using the FEA methods described herein, by the motor (or drive equipment manufacturer. In such cases land at the discretion of the party having unit responsibilty, impact tests shall be performed on vertical motors or other drive equipment to verity the reed frequency value obtained analytically, o°if different, to determine if design adjustments are necessary. + Motor or drive critical speed. + Coupling or remote shatting effects + Effects of iquid external to the pump at the highest and lowest liquid levels (wet pit pumps only). + Effects of internal liquid being pumped including Lomakin effect, + Any miscellaneous damping effects. + Effects of bearings, wear rings, and close-clear: ance bushings. Hydraulic nsitute Standards, Copyright © 1997-2005, Al Rahs Reserved 3HI Pump Vibration — 2000 + The system shall be analyzed in more than one ‘axis, f component characteristics in one axis differ significantly from another. A complete dynamic analysis report should be submit- ted to the purchaser in the pre-construction stage, and. should include the following information’ a) Computer program used ) Assumptions made in order to model the system ©) Schematic diagram of the model depicting nodes and elements d) Input data consisting of node coordinates, ele- ment types, material properties, element charac- teristics, element connectivities, and specified displacements ) Any supporting calculations and data 1) Motor or other drive reed frequency data (or dynamic analysis, if required) g) Motor or other drive critical speed data, by the manufacturer h)_ Shatting critical speed data, by the manufacturer i) Anticipated excitation frequencies |) Analysis results including all significant natural fre- {quencies and a description of the corresponding mode shapes k) An interference diagram depicting a plot of excita- tion frequency versus operating speed, with the following information’ 1) A plot of all excitation sources that are mult ples of the operating speed, propery labeled 2) A plot of natural frequencies versus operating 3) Areas of interference of 1) and 2) described above clearly identified, or the lack of likely resonance indicated |) Interpretation of the results and recommenda- tions, including recommendations for appropriate corrective action 9.6.4.2.3 Torsional critical speed ‘The term torsional ential speeds of centrifugal pumps and associated drive equipment, refers to the speed of 2 pump rotor or related rotating system that corre sponds to 2 resonant frequency of torsional vibration of the rotating system (a resonant frequency is one at which resonance occurs). Torsional critical speeds are associated with torsional or angular deflection of the: rotor and are not to be confused with lateral critical speeds associated with lateral deflection. The two are separate entities. A given rotor or rotating system may possess more than one torsional resonant frequency and, therefore, torsional critical speed. The lowest [re- quency, which produces the “first mode shape" and “frst critical speed,” is in general of the most concern for most types of equipment. Determination of torsional critical speed can be impor- tant in pumps and associated rotating equipment because if a torsional critical speed or resonant fre- quency is close to an exciting frequency, small excita tion forces can be greatly amplified. The resulting stresses and angular deflections can cause premature equipment failure, Examples of sources of torsional excitation include variable frequency motor torque pul. sations, combustion engine torque spikes, and impel- ler vane pass, particularly for impellers with only two or three vanes, The calculation of the first torsional crit- cal speed is fairly simple for simple rotor systems Figure 9.6.4.3 represents a simple horizontal centrifu- gal pump system consisting of a pump, motor, con- necting shafts, and flexible coupling, with associated rotational inertia and torsional sifiness values. speed ouning Motor Pump Figure 9.6.4.3 — Simple horizontal centrifugal pump system 4 Hydra Instute Standards, Copyright © 1897-2005, Al Rights ReservedThe equivalent torsional stifitess of the system, Ke, is calculated by the equation: The torsional critical speed may be calculated as: We, ~ $0] Ketdm + dray® 2n. Javdp Where: Ne; = torsional critical speed (rpm) Ke, Km, Ke, Kp = torsional spring constants (kN-mirad [Ib-in Jrad}) @ = gravitational constant (9.81 m/sec? [386 inJsec*)) Jm, Jp = rotational inertias (kg-m? {Ib-in)), also known as WK? or WR? W> = weight (kg {Ibs)) K or R = radius of gyration (mm [in-4}) The rotational inertia of the liquid within the impeller is considered by adding the liquid rotational inertia to that of the pump impeller, resulting in the “wet” impel- ler rotational inertia. Other equations are available or may be derived for simple systems, however, manual calculation methods become prohibitive as the number of components in the rotating system increases (e.g., multiple impellers, shatts, couplings, etc.). Calculation of K values can become cumbersome with stepped shafts. Also, evalu- ation of modes other than the first mode is dificult. For more complicated calculations, various computerized methods are used. Application-specific computer pro- grams are available that determine torsional critical speeds for specific cases, in particular engine-driven equipment. Engine-driven systems are often analyzed by the engine supplier or others familiar with the inter- rnal design of the engine, authorized by the vendor having system responsibly. The method of calculat- ing torsional critical speed, as well as the extent of the analysis, should be agreed upon by the supplier and the purchaser. Finite element analysis methods (Section 9.6.4.3.4), provide a specication by which torsional crit cal speeds may be determined using these methods. HI Pump Vibration — 2000 Rotating systems are typically designed so that the torsional critical speeds are 10% to 25% away from the rotating speed. Many systom torsional resonance problems are typically solved by a calculated change in the torsional stiffness of one system component, typically the coupling between the pump and its driver. Experimental testing to determine torsional critical speeds) is often possible, however, sometimes at con- siderable expense relative to the value of the equip ment. Special equipment is needed to measure torsional vibration. Methods of experimental testing (if required) should be agreed upon by the supplier and the purchaser. The vendor having unit responsibility shall interpret the results of any such tests and make recommendations regarding any required corrective action. 9.6.4.2.4 Torsional dynamic analysis (optional) The torsional dynamic analysis may be performed using finite element analysis methods commonly uti lized in commercially available FEA computer codes or ther proven computer programs. The model shall consider the characteristics af any structural compo- nent or mass as necessary to allow the determination of the torsional critical speed characteristics of the equipment, including the pump and driver rotational inertias, pump and driver shaft rigidities and inertias and rigidities of all other rotating equipment in the drive train between the pump and driver. Torsional data pertaining to couplings or intermediate shatting shall be incorporated in the model. Acomplete dynamic analysis report should be submit- ted to the purchaser in the pre-construction stage, and should include the following information: a) Computer program used b) Assumptions made to model the system ©) Schematic diagram of the model depicting nodes land elements. 4) Input data consisting of node coordinates, ele ‘ment types, material properties, element charac: teristics, element connectivities, and specified displacements ©), Any supporting calculations and data 1) Pump and drive equipment torsional data by the respective manufacturers Hycraulc Institute Standards, Copyright® 1987-2005, Al Rights Reserved 5HI Pump Vibration — 2000 {9}. Shafting or coupling torsional data by its ‘manufacturer h) Anticipated exciting frequencies |) Analysis results including all significant torsional critical speeds and a description of the corre- sponding mode shapes |) Aniinterference diagram depicting a plot of excita- tion frequency versus operating speed, with the following information’ 1) A plot of all excitation sources that are multi ples of the operating speed, properly labeled 2). A pilot of torsional natural frequencies versus operating speed 3) Areas of interference of 1) and 2) described above clearly identified, or the lack of likely resonance indicated k) Interpretation of results and recommendations, including recommendations for appropriate corec- tive action 9.6.4.2.5 Structure lateral vibration ‘Structural vibration occurs in both horizontal and vert cal pumps. However, vertical structures are particu- larly vulnerable to resonance because the natural frequencies of these structures are inherently lower than those involving horizontal pumps and are nearer to the operating speed of the equipment. A vertical structure, which may consist of a pump only, a motor only, a pump with drive equipment (motor, ‘gear, etc.) directly mounted, and associated bases or supports can be idealized as a vertical cantilever beam, anchored at one end and free to move at the other end. Each such structure possesses its own nat- ural frequencies. Each component ofthe structure has its own contribution to the natural frequency character- istics of the structure Determination of any structure natural frequency is important because if a structure natural frequency is numerically close to an operating speed or other exci ing frequency, a state of resonance exists wherein any excitation force can be greatly amplified. The tre- quency at which resonance occurs is known as a reso- nant frequency. At a resonant frequency, the resulting stresses and deflections can cause premature equip- ‘ment failure 9.6.4.2.5.1 Vertical dry pit pumps The lowest (or first) natural frequency of a vertical structure, ifidealized as a vertical cantilever beam with ‘concentrated mass, can be illustrated using the fol lowing equation: Metric: fy = 0.0299 / dyy® US Units: fy = 187.7 /dgy® Where: fy = natural frequency (cycles/min) dy = static deflection (mm [in of the center of ‘gravity of the structure caused by a force ‘equal to the weight of the structure applied horizontally at the center of gravity of the structure being analyzed The natural frequency of a structure is sometimes called the reed frequency, due to its similarity to the oscillation of a reed. The natural frequency depends Con the mass and stifness of the structure. From the equation it can be seen that the natural fre- quency will decrease as deflection increases. Deflec- tion will increase as structure weight increases, as Structure rigidly decreases, and as the distance of the. Center of gravity from the support increases. Conversely, the natural frequency will increase as deflection decreases. Deflection will decrease as the structure weight decreases, as. structure rigidity Increases, and as the distance of the center of gravity from the support decreases, There will be higher natural frequencies not repre- sented by this simple model, which correspond to other modes of vibration, These natural frequencies are often far remaved from the operating speed so as tonot be a problem, 9.6.4.2.5.2 Vertical wet pit pumps The above discussion concerning dry pit pumps applies to wet pit pumps as well, with some difer- ences. Wet pit pumps usually have a structure above the base and a structure below the base, each with its ‘own natural frequency characteristics. The natural fre- quency of the structure below the baseplate will be affected by the water around the submerged pump col umn, and therefore the effects of high and low water 6 Hydra insttute Standards, Copytight © 1097-2005, Al Rights ReservedBERR ERE EEE EEE EEE ERE ER REE ESE levels must be considered. If an outer containment is sed, it must be included in the analysis of vibration 9.6.4.2.6 Structure dynamic analysis (optional) In the pre-construction stage it is desirable to verify that a structure has no natural frequency near an oper- ating speed. This is particularly important for variable- ‘speed installations, due to a range of speeds that must be avoided. Calculation of the first and other natural frequencies and the corresponding mode shapes can become quite complicated. Natural frequency data for each component of the structure must be incorporated Into the analysis, and as such, adequate natural fre- quency data for each structure component (motor, for example) must be obtained and included by the party having responsibiliy for the pump and driver assembly. The simple modal in Section 9.6.4.3.5 is often inade- uate for the purpose of analysis because of the dif culty of determining the location and deflection of the center of gravity in such structures. The model is included here for ilustration purposes only, and not for purposes of analysis. ‘Computational methods using application-specific pro- grams or general-purpose finite element analysis (FEA) programs may be required to produce adequate results, Even so, the actual distribution of the structure mass and stiffness can be diffcul, effecting the accu- racy of the calculation. Typical sources of inaccuracies include system nonlinearities and joint looseness, A sample specification involving FEA methods is pro- vided in Section 9.6.4.3.4 The method of calculating structure natural frequency, the extent of the analysis, and the margins of safety to be used should be agreed upon by the purchaser and the manufacturer. Margins of safety between structure natural frequency and operating speed obtained by calculation vary within the industry. Typical margins are 25%, particu- larly in structures involving multiple components and! oF multiple bolted joints, however some application- specific industry specifications allow margins as low 28 10%. Margins of safety on the order of 25% may be impractical in certain applications, such as those hav- ing several closely-spaced natural frequencies operat- ing at variable speed, with multiple excitation sources, In some applications, low margins of safety between ‘aperating speed and actual (not calculated) natural frequency can result in successful operation, depend- ing on the levels of excitation and damping present. HI Pump Vibration — 2000 The natural frequency of a structure is unique to a 4iven installation and may not be measurable in any other mounting, such as in a shop test. Normally, a pump unit in a shop test with a non-rigid support will have different natural frequencies than when itis rig idly installed in a typical permanent system. The first natural frequency or reed frequency may be determined by exciting the structure with a variable speed frequency exciter. A vibration transducer together with a vibration analyzer is used to determine the resonant frequency as exciter frequency is varied. Alternately, the resonant frequency may be obtained approximately by impact testing. In this technique, the structure is excited by striking with a soft hammer (hard-taced steel hammers are not recommended). A 200 200 mm (4 ¥ 4 inch) piece of timber 1 meter (3. feet) long also works well. This will cause the structure to vibrate at its resonant frequency. A single-channel vibration analyzer may be used to measure the fre- quency ofthis vibration. A more elaborate and expen- sive impact test, when justified, may be performed using a multichannel analyzer with an instrumented hammer or shaker to obtain all the modal characteris- ties (natural frequency, made shape, and damping) of the structure and to ensure that the measured response is caused by the forced input, and not effected by background noise or electrical interference, Structures are not generally symmetrical and may therefore have different natural frequencies in diferent axes, Any axis of significant mechanical dissymmetry should be checked. The axis parallel to the pump dis- ccharge piping and the axis perpendicular to the piping are axes commonly having significant mechanical dis symmetry for pump structures, which should be checked The party having responsibilty for the pump and driver assembly shall interpret the results of any such tests and make recommendations regarding any corrective action. 9.6.4.3 Vibration measurements 9.6.4.3.1 bration instruments, The primary measure of vibration is velocity, unfiltered, in milimeters (inches) per second RMS. “The most appropriate method of measurement is elec: tronic velocity or acceleration measurements takon at various frequencies. Where the values are significant, these measurements are integrated in an appropriate Hydrauleinetute Standards, Copyright© 1997-2005, Rgh's Reserved 7HI Pump Vibration — 2000 ‘electronic circuit to determine overall vibration in mili meters (inches) per second, Direct measurement of velocity or acceleration filtered to unique frequencies is not intended by this guide. ‘Such secondary measurements, including complete frequency analysis, are useful in diagnosing vibration problems. It should be noted that relatively high velocity or accel- eration readings at high frequencies result in very small displacement values. 9.6.4.3.2 Location of vibration probes ‘The vibration probes should be located on both bear- ing housings of horizontal spit case or double case pumps, near the upper or outer bearing of end suction pumps and near the upper motor bearing of vertical Bumps with close-coupled motors. Figures 9.6.4.4 through 9.6.4.14 illustrate the location in greater detail. Probes must not be located on flexi- ble panel or cylinder walls, such as on motor end cov- es of vertical pumps. Such covers should be removed to allow measurements on a stif part of the machine. Displacement measurements of the pump sha, rela- tive to the bearing journals, are sometimes taken with proximity probes supported on the bearing housing, ‘Such measurements are primarily useful as analytical tools, since they provide information on phase angles, vibration modes. and frequencies. However, amplt tudes are influenced heavily by the bearing clearance and cil film and are normally larger than the housing vibration. They are not necessarily in conformance with the guidelines in Figures 9.6.4.4 through 9.6.4.14, 9.6.4.4 Allowable pump field vibration values The vibration values shown in Figures 9.6.4.4 through 2.6.4.14 are for unfitered RMS velocity readings. These values assume the following conditions: * Operation under steady state conditions at the rated speed +/- 10% + No entrained air or gas and adequate NPSH mar- gin (see ANSI/HI 9.6.1-1998, Centrifugal and Ver- tical Pumps ~ NPSH Margin) Operation within the pump's Preferred Operating Region, POR, as recommended by the pump man- ufacturer (see ANSI/HI 9.6.3-1997, Centrifugal and Vertical Pumps ~ Allowable Operation Region) + Pump must be installed so that shaft alignment and flange loads are kept in accordance with the ‘manufacturers’ recommendations + Vibration level recorded is the maximum of mea- surements taken in each of three planes; vertical, horizontal or axial; measured as shown in Figures 9.6.4.4109.6.4.14 + Pump intakes (wet wells) shall be properly designed in accordance with ANSWHI 9.8-1998, Pump Intake Design The values in Figures 9.6.4.4 to 9,6.4.14 are not appli- cable to factory or laboratory acceptance tests. Experi- ence has shown that vibration levels measured on temporary factory setups may be as much as two times higher than those obtained in the fel. The field vibration acceptance levels are applicable to the pump when operating anywhere within the pre- ferred operating range (see ANSIMHI 9.6.3-1997), When entering the charts, the abscissa refers to the ower the pump is drawing at the time the vibration measurement is made, To illustrate, assume a pump was sold for rated duty point of 225 mM (1000 gpm), with a preferred operating range of 110 to 250 mh (500 to 1100 gpm). At the rated duty point, the pump would draw 37 kW (50 hp). A vibration reading is taken in the field when the pump is operating at 180 min (800 gpm), at which time itis drawing 34 KW (45 hp). To evaluate the vibration level, one would enter the appropriate chart at 34 KW (45 hp) and read the acceptable field vibration level. If ater the pump rate of flow increases to 250 m’m (1100 gpm), where the ower consumed is 45 kW (60 hp), then one would enter the appropriate chart at 45 kWV (60 hp). I the rate of flow were less than 110 m°sh (500 gpm), the vibra- tion acceptance standards would not be applicable because the pump was operating below the minimum of preferred operating range. These vibration values are to be used as a general acceptance guide with the understanding that vibration levels in excess of these values may be acceptable by mutual agreement if they show no continued increase with time and there is no indication of damage, such {8 an increase in bearing clearance or noise level 8 Hydra Insttute Standards. Copyight © 1957-2005, All Rights ReservedHI Pump Vibration — 2000 15 75 750 Figure 9.6.4.4 — End suction foot mounted-ANSI B73.1, B73.3, B73.5, Hydraulic institute Standavds. Copyight © 1967-2005, All Rights Reserved 9HI Pump Vibration — 2000 Figure 9.6.4.5 — Vertical in-line, separately coupled, ANSI 873.2 10 Hydraute lnstute Standards, Copyright© 1997-2005, A Rights ReservedHI Pump Vibration — 2000 Figure 9.6.4.6 — End suction, close coupled horizontal and vertical in-lineHI Pump Vibration — 2000 Figure 9.6.4.7 — End suction, frame mounted 3 : i 2 3 8 :Rg ESB SB SBS SB SB SB SS SB S82 82 82 32 32 32 2 2 3) HI Pump Vibration — 2000 Figure 9.6.4.8 — End suction, centerline support, API 610 Hydrate Insttute Standards. Copyright© 1997-2006, All Rghis Reserves 13HE SeBE SBE SE SE SB SB SB SB SBE SB SB SB SB SB SB SS HI Pump Vibration — 2000 ne — kW Figure 9.6.4.9 — End suction, paper stock 4 Hydra Instute Standards, Copyight © 1887-2008, Al Rights Reserved| HI Pump Vibration BBE S SB SB 2 Figure 9.6.4.10 — End suction, solids handling, horizontal and vertical Hydraulic Insitute Standards, Cpyight © 1967-2005, Al Rights Reserved 2000HI Pump Vibration — 2000 Figure 9.6.4.11 — End suction, hard metal and rubber-lined horizontal and vertical Hydraulic Insttute Standards, Copyright© 1997-2005,HI Pump Vibration — 2000 2 408 A 1 203 Figure 9.6.4.12 — Between bearing, single and multistage Hydrauic Inetute Standards, Copyright© 1907-2006, Al Rights Recerved v7HI Pump Vibration — 2000 18 018 DISCHARGE HEAD | et a I Figure 9.6.4.13 — Vertical turbine, mixed flow and propeller type Hydraute Insitute Standards, Copyright © 1997-2006, Al ights ReservedHI Pump Vibration — 2000 07s 7: Ee 750 2208 Input power @ test condtions — KW Noror Input power @ ts! conaitons — BHP | onsenarce oiscwaRor nao el Figure 9.6.4.14 — Vertical turbine, short set pumps, assembled for shipment by the manufacturer Hydraulic nstute Standards. Copyright © 1997-2005, AllRighls Reserved 19HI Pump Vibration — 2000 9.6.4.5 Factors affecting vibration 9.6.4.5.1 Unbalance of rotating parts and rotor balancing Unbalance of the pump rotor can generate high unbal- ance forces, result in excessive bearing and shaft foading and induce high vibration levels. The balancing method and the residual unbalance limits are as described in this section, Component balance shall be single-plane spin bal~ ance to ISO 1940iI:1986 balance quality grade G6.3, (See Figure 9.6.4.15 A and 8). When the ratio of the largest outside diameter of the component divided by the distance between the correction planes is less than six, @ two plane balance may be required. Other grades may be used if agreed upon by the user and ‘manufacturer. NOTE: In the specific case of impellers, the width is measured at the periphery, including the thickness. cof any shrouds but not the back vane. Figure 9.64.15 is used by entering the graph at the maximum expected service speed, such as 3000 rpm, and reading the acceptable residual unbalance as (021 kg ~ mmikg (.8502 - in/oz). Multiply this number by the rotor weight in kg (oz) and the result is the allowable unbalance of the rotor in kg ~ mm (oz = in). Balancing machine sensitivity shall be adequate for the part to be balanced. This means that the machine is capable of measuring unbalanced levels to one- tenth of the maximum residual unbalance allowed by the balance quality grade selected for the component boing balanced. Balancing machines are capable of measuring unbal- ance independent of its speed. When the value for allowable unbalance is determined from figure 9.64.15, i is not necessary to operate the balancing machine at the same speed as the pump speed. Balance machines shall be calibrated as recom- mended by its manufacturer. When specified, calibra- tion shall be done just prior to balancing, Pertinent aspects: — The balance grade specified in the above standard yields a level or residual unbalance in rotating ‘components consistent with clearance fits between the impeller and shaft, Rotaling assembly balance is recommended when practical if tighter {quality grades (e.g., G2.5 or G1 0) are desired. For those instances where non-clearance fits are applicable, agreement between the manufacturer ‘and purchaser should be reached if a different ‘component balance quality grade Is desired; — Balance machine sensitivity is a function of the ratio of the weight of the part tothe weight rating of the machine, (Above 100%, one must check with the manufacturer of the balancing machine.) AS an example, a 45-kg (100-b) rated machine may pro- vide adequate sensitivity and accuracy for a 4.5- kg (10-1) part, but @ 9kg (20-1b) rated machine ‘would be much more suited for the task; and a 1.5- kg (3-10) part may not balance at all on the 45-kg (100-1b) machine to the quality grade required. Following are guidelines for the quality of balance pro- cedure, equipment, tooling and rotor geometry, giving both users and manufacturers a common ground for discussing these issues that have been leaned through experience. 41) Inherent balance andlor runout in balancer drive or balancing arbor. ‘The balancer drive may be checked by periodically rotating the drive splines 180 degrees after a part has been balanced and checking the residual tunbalanee. it should be within 10% of the original Uunbalance. Runout in the balancing arbor should be checked when assembled in the balancer. It should be no more than .025 mm (.001 in.) total indicator movement 2). Keysikeyway geometry errors. ‘Special care must be taken to ensure that keys and keyways in balancing arbors are dimensionally identical and the same weight to those in the assembled rotor. Like the arbor, they should be of hardened tool steel to resist error introduced through wear, 20 Hydraulcnsttute Standards, Copyright © 1997-2006, Allah's ReservedHI Pump Vibration — 2000 ers/kiogram of rotor mass ‘606 rotation (7pm) x 10 Figure 9.6.4.15A — Allowable residual unbalance in pump impellers ~ grade G6.3 (metric) Hydeaulcnstute Standards. Copyright © 1997-2005, All Rants Reserved a— ee ee 3) Excessive looseness between impeller hub and balancing arbor. The following guidelines are suggested for maxi- mum looseness between balancing arbor and Impeller. At no time should this looseness be ‘greater than that found on the assembled rotor; Diameter impeller Maximum looseness, hub bore (diametral) (51800 cpm) (1800 < 3600 rpm) o— 038 mm 038 mm 38mm (1.499in.) —_(0018in.)__(.0075in.) 38-1 mm (1.5 in.)— 054 mm 038 mm 50.8 mm (1.999 in.) (,0020in.) (0018 in.) 50.9 mm (20in.) 065 mm 038 mm — above (.0025 in.) (.0018 in.) NOTE: For those impellers to be run at 3,800 rpm and hav: ing a straight racial clearance with the shaft, the maximum loaseness speciied, while a practical imi, may not assure a (G63 residual unbalance when removed from the balance arbor ané mounted on the pump shaft. In this case, a G10 Spectfeation on impeller balance ora rtating assembly bal lance is recommended, 4) Removal or addition of material Material removal: This should be done in a man- nner that spreads the balance correction as evenly as possible over the surface, if shroud is used, the thickness removed should be no more than 1/3 of the original, and the subsequent finish should be ‘equal o the remainder of the shroud. If the impeller vane is used for balance correction, no more than 1/4 ff the vane thickness should be removed, always. ‘rom the low-pressure side. Removal by driling and! for end milling should follow the same thickness guidelines, with appropriate consideration to mini ‘izing flow discontinuities. Material addition: Sometimes for very large rotors with large amounts of unbalance it becomes desir able to add material so that the shroud/vane thick- ness guidelines are not violated. This is permissible as long as impeller finish and discontinuities to flow ‘are not radical and the method of material adaion is consistent with requirements. for mechanical integrty and material properties of the impeller! component for the intended/specified service. HI Pump Vibration — 2000 The magnitude of pump vibration caused by rotor unbalance is also influenced by the mass of the pump ‘casing. As the size of the pump casing increases, the magnitude of the vibration measured in the bearing housings decreases. Unbalance of the driver, especially when it is mounted ‘on the pump, and unbalance of the coupling will have the same effect as pump rotor unbalance and require the same treatment, Field balancing is an acceptable means for meeting guaranteed field vibration limits. it is sometimes. impractical to balance a pump driver assembly in the shop because of the size, speed, thermal conditions of the assembly, etc. Furthermore, itis found that a shop- balanced assembly may produce different vibration levels in the field because of the dissimilarities between shop and field foundation, job or test driver, Piping, etc. 9.6.4.5.2 Natural frequency and resonance Operation of a pump at a rotational speed near or at vane-passing frequency or one of the lower natural fre- quencies of the structure can result in a resonant con- dition, Theoretically, the vibration levels could become infinite, but the presence of dampening generally limits the levels to several times that shown on Figures 96.4.4 through 8.64.14. Pump manufacturers can calculate or determine by test the natural frequency of the pump assembly. How- ever, in afield installation, the vibrating structure com- prises, in addition to the pump, the foundation, the ‘mounting, the piping and its supports, and may include the driver and flexible shafting. The natural frequency of the vibrating structure is determined by the stiffness of the total structure and by its equivalent mass. It may therefore differ significantly from the natural frequency of the pump alone. In the absence of any specific information, the pump manufacturer will assume that the piping is installed figidly and anchored close to the pump connections ‘and that the hold-down bolts are securely embedded in a concrete foundation of infinite mass and riidiy. The system designer must give this proper consider- ation and must ensure that the natural frequency of the system does not fall within the pump operating speed range. He also must be aware of the much lower stiffness of fabricated system structures rela- live to conerete and the problems associated with calculating stiffness of unconventional and composite Fydrautc Instole Standards, Copyight © 1997-2006, All Righls Reserves 23HI Pump Vibration — 2000 structures. This is especially true in the case of verti- cally mounted pumps in applications where variable speed drives are employed. Horizontal pumps are rel- atively rigid with a resulting high natural frequency, and ‘occasions where resonance problems arise are rare. 9.6.4.5.2.1. Hydraulic resonance in piping Vibration problems can be caused by hydraulic reso- nance of the liquid within the pump/piping system. When normal pump-induced pressure pulsations are reflected by the piping systems and added in phase to the source pulse, the amplitude of the pulse is magnified. Resulting high pressures can ultimately cause mechanical fatigue failures in either the piping or the pump components, can generate unacceptable noise levels and can cause vibration of the system components. Experience has shown that the following measures may prove effective in correcting hydraulic resonance: a) alter the resonant piping; ) change the pump speed, ©) change the internal design characteristics of the pump: d)_ insert a pulsation damper on the pump/piping system. Modifications to the pump, piping or supporting struc- tures that do not change the pulsation response (hydraulic natural frequency) of the pump/piping sys- tem will not correct the resonant condition. 9.6.45. Hydraulic disturbances Vibration is atways caused by a driving force. Hydraulic disturbances in the pump may generate this force. Fol lowing is alist of some typical hydraulic disturbances: a) hydraulic forces produced between the impeller vanes and the volute cutwater or difuser at vane- passing frequency; recirculation and radial forces at low flows; ) fluid separation at high flows 4d) cavitation due to NPSH problems; 8) flow disturbances in the pump intake due to improper intake design: f) air entrainment or aeration of the liquid; 9) hydraulic resonance in the piping; h). solids contained in the liquids, such as sewage, impacting in the pump and causing momentary imbalance, or wedged in the impeller and causing ‘continuous imbalance. ‘The pump manufacturer should provide for item “a ‘above in the pump design and establish limits for iow low. The system designer is responsible for giving due consideration to the remaining items, 9.6.4.5.4 Other mechanical problems Misalignment of the shafting, damaged bearing, bent shafts, inadequate piping supports and expansion joints without tie rods are also frequent causes of vibration 9.6.4.5.5 Effect of rigidity ‘The amplitude of the vibration resulting from a given diving force is related to the rigidity of the vibrating structure. For example, a conventional horizontal split case pump is much more rigid than a vertically mounted sewage pump. The same forcing levels of any of the above sources could be expected to result in signifi ‘cantly higher vibration amplitude levels on the sewage ump bearing frame than on the split case. Likewise, a short vertical sewage pump would be more rigid than a tall one, and lower vibration amplitudes would be expected. The higher flexibility of the taller structure, however, results in higher vibration ampli- tudes for the same effect on life of the components. 26 Hysrou Institute Standards, Copyright © 1957-2005, A Rights ReservedHI Pump Vibration — 2000 Appendix A Bibliography Lobanotf and Ross, Centrifugal Pumps, Design and Application, Gulf Publishing Co,, 1985. Karassik, Krutzsch, Frazer and Messina, Editors; Pump Handbook, 1st Ed., McGraw Hil, 1976, pp:9) 90. Kovats, Design and Performance of Centrifugal and Axial Flow Pumps and Compressors, The Macmillan Co,, 1964; pp:456-7. Kovats, “Vibration of Vertical Pumps", Paper No. 61- Hyd-10, Transactions of the ASME, Journal of Engi- neering for Power. Dicmas, Vertical Turbine, Mived Flow, and Propeller Pumps, MeGraw Hill, 978; pp.98. ydrauteIstte Standards, Copyright © 1997-2005, AX Rights Reserved 25