Jtam-A4.dvi JOURNAL OF THEORETICAL AND APPLIED MECHANICS 54, 1, pp. 251-262, Warsaw 2016 DOI: 10.15632/jtam-pl.54.1.251 STRENGTH CALCULATIONS OF AN ELEMENT COMPENSATING CIRCUMFERENTIAL BACKLASH IN THE EXTERNAL GEAR PUMP Piotr Osiński, Grzegorz Chruścielski Wroclaw University of Technology, Faculty of Mechanical Engineering, Wrocław, Poland e-mail: piotr.osinski@pwr.edu.pl; grzegorz.chruscielski@pwr.edu.pl The aim of the article is to discuss the issue of strength of the circumferential backlash compensating beam in a high-efficiency gear pump. Three geometric versions of the com- pensating beam structure differing in the wrapping angle are analyzed. The mechanical model is solved assuming the curved beam model supported at the contact points between the beam and the gear teeth. The assumedmechanical structure is statically indeterminate. In order to determine the reactions in the supports and in the beam fixing, the Menabrei and Castigliano theorems are used. Based on analytical calculation results, the cause of compensation structure damage during experimental tests of prototype units is identified and the most favorable variant of beam structure, from the mechanical strength point of view, is determined. Keywords: external gear pump, strength calculation, circumferential backlash compensation 1. Introduction The efficiency of a gear pump is to a large extent determined by clearances between the gears and the elements limiting the displacement chamber volume of the pump (Chrobot et al., 1997; Judin, 1958; Kollek, 1996, 2004; Osiński et al., 2013; Osiński and Kollek, 2007; Ragunathan and Manoharan, 2012; Vacca and Guidetti, 2011; Wang et al., 2011). Two types of clearance, i.e. radial clearance and frontal clearance, are distinguished. The former is also referred to as circumferential clearance (backlash). The circumferential gap is formed by the surface of casing concavities and that of the cylinder with the radius of the addendum circle of toothed displacement elements. The gap is not constant along the whole circumference the gears often move within the bearing slackness limits towards the suction space. In conventional pumps without radial backlash compensation, the circumferential gap assumes the shape of a crescent widening towards the delivery side. In such a pumpdesign, the gap ranges from 0.01 to 0.3mm. The circumferential clearance is a gap with one fixed wall and one movable wall moving in the direction opposite to that of the pressure drop.This is an advantageous configuration as a result of the rotational motion of the gear the liquid is lifted from the suction space to the delivery space, thereby reducing the leakages due to the pressure difference between the gap ends (Singal et al., 2009; Stryczek, 1995). Frontal clearances have the shape of a ring limited by the diameter of the dedendum circle and that of the gear shaft. Most of the leakages are used to cool and lubricate the bearings. After they pass through the bearings, the leakages are directed via special grooves to the suction chamber, but someof the volume losses pass directly through the gap into the suction space.The frontal clearance values are by one order of magnitude lower than the radial clearance values. In typical pumps, frontal clearances are in a range of 0.01-0.05mm. The recommended clearance also depends on the pump size. Lower values are recommended for units with a lower specific output. 252 P. Osiński, G. Chruścielski Nowadays backlash compensation is used to improve the efficiency of gear pumps (Kollek and Radziwanowska, 2015). In addition, this treatment contributes to better running in of the interacting parts andmaintenance of a constant gap despite thewear of the parts. In the curren- tly produced pumps, mainly the axial backlash compensation is carried out. However, if higher efficiency is required, it is necessary to compensate also the circumferential gap. The latter can be compensated in two ways. One way consists in compensating backlash locally along a short distance. It is further referred to as radial backlash compensation (Fig. 1a). Another method, developed by the authors, consists in ensuring a constant gap along the whole circumference (Osiński, 2012a,b; Wiczkowski, 2012).The method is called circumferential backlash compensa- tion (Fig. 1c). Fig. 1. Schematic showing the way of sealing gears along circumference plus graph of circumferential pressuremeasured in the gear root: (a) pump with radial backlash compensation, (b) pump with axial backlash compensation, (c) pump with circumferential backlash compensation (Osiński et al., 2012a) The influence of the applied compensation on the overall efficiency of the pump is shown in Fig. 2. The comparative diagram is based on specifications found in themanufacturer catalogues of Bosch, Casappa, Marzocchi, Hamworthy, Hidroirma, Orsta, Parker, PZL-Hydral, Rexroth, WPH, VPS and on the authors’ own studies of prototype pumps with circumferential backlash compensation (Osiński, 2013). It appears from the diagram that the application of different methods of backlash compen- sation considerably increases the efficiency and working pressure of gear pumps. The currently produced pumps reach working pressures as high as 32MPa. The innovative circumferential backlash compensation method enables one to increase the pressures by nearly 20%, i.e. to the level of 40MPa. The increasing of the internal tightness also makes it possible to increase the total efficiency by about 5% on average. 2. Circumferential compensation structure There are three versions of the displacement pump structure with a compensating pressure chamber (Osiński, 2013;Osińskiet al., 2012), differing in thedesignof the compensatingchamber Strength calculations of an element compensating circumferential backlash... 253 Fig. 2. Comparison of total efficiency ηc of gear pumps with regard to forcing pressure pt and backlash compensation (based on catalogues of majormanufacturers + own research): 1 – without compensation, 2 – with axial compensation, 3 – with axial and radial compensation, 4 – with axial and circumferential compensation (Fig. 3). In this pump, two interacting gears perform rotations in the directionsmarked inFig. 3, forcing the working liquid (oil) from the suction chamber on the left side of the pump (Fig. 3) through the inter tooth spaces into the delivery chamber on the right side of the pump. Fig. 3. Schematic of the displacement pumpwith different compensating beam designs: (a) beamwith a wrapping angleϕ0 =102 ◦, (b) ϕ0 =132 ◦, (c) ϕ0 =169 ◦ Experimental studies of prototypes of such pumps have shown that during operation under heavy loads (at pressures p above 20MPa) the beam closing the compensating pressure chamber is susceptible to failure in the place of its fixing because of too small beam thickness. The aim of the calculations presented in this paper is to determine (from the strength condition) the minimum thickness h in the fixed cross section of the beam ensuring that the stresses in this cross section are carried. Three geometric versions of the compensating beam structure have been designed. Themost optimal version will be selected on the basis of theoretical calculations and experiments. The versions differ in the beam length, i.e. its wrapping angle, and so in the number of teeth inter- acting with the beam. For the statical analysis, a fixed gear position in which one of the teeth is in contact with the beam fixing cross section is assumed in each of the cases. The versions include: a) a beamwith a wrapping angle ϕ0 =102 ◦, interacting with three teeth of the gear, one of which is in contact with the beam fixing cross section (Fig. 3a); b) a beam with a wrapping angle ϕ0 = 132 ◦, interacting with four teeth of the gear, one of which is in contact with the beam fixing cross section (Fig. 3b); c) a beam with a wrapping angle ϕ0 = 169 ◦, interacting with five teeth of the gear, one of which is in contact with the beam fixing cross section (Fig. 3c). 254 P. Osiński, G. Chruścielski 3. Static calculations for the pressure chamber beam 3.1. Beam with a wrapping angle ϕ0 =102 ◦ 3.1.1. Beam geometry and loading diagram During the operation of the pump, the compensating chamber beam is loaded from the out- side with compensating pressure p2 constant along the whole length of the beam, and from the insidewith working pressure p1 (Fig. 4a). The pressure p1 decreases in the successive intertooth spaces from the initial value p1p = p2 at the inlet to the compensating chamber up to end value p1k = 0.5p2 in the tooth space at the beam fixing. Thus the pressure difference ∆p= p2−p1k constitutes a linearly variable continuous load q(ϕ) for the beam, whose initial value is q(ϕ = 0) = 0 and its end value is q(ϕ = 2ϕ1 +α) = q0 = ∆pb, where b is the beam width (Fig. 4b). Then the beam load can be reduced to a flat system. Fig. 4. (a) Static diagram of the beamwith the wrapping angle of 102◦, loaded with working pressure p1 and compensating pressure p2, (b) diagram after introduction of continuous load replacing the action of pressures p1 and p2 Asa result of thedifferencebetween thepressurep2 andp1, thebeamcomes into contactwith the pump gear teeth, constituting movable supports of the beam, in points A and B (Fig. 2b). Respective reactions RA and RB and friction forces TA and TB, whose sense is consistent with the direction of the rotational motion of the pump gear, occur in the supports. Two reactions: RCx andRCy and fixing moment MC occur in the beam fixing place (pointC). The directions of reactions RCx andRCy correspond to the adopted flat reference system (xy) whose origin is in the centre of gravity of the fixed cross section and which is connected with the normal and tangent direction of this cross section (Fig. 4b). The angles ϕ1 (marked in Fig. 4) between the central surfaces of the teeth amount to ϕ1 = 360 ◦/10 = 36◦ (the pump gears have 10 teeth) while the complementary angle between pointA and the beginning of the beam amounts to α=30◦. The beam width (in the direction perpendicular to the load surface) is constant and amounts to b=26.3mm. 3.1.2. Solutions for beam static load system For the assumed beam loading diagram (Fig. 4b), the static equilibrium equations have the form ∑ Px =RCx−Qx+RBx+RAx+TC +TBx+TAx =0 ∑ Py =RCy−Qy +RBy +RAy−TBy −TAy =0 ∑ MC =MC +RBr sinϕ1+RAr sin2ϕ1−TB(r−rcosϕ1)−TA(r−rcos2ϕ1) −Qr sin(ϕc/3)= 0 (3.1) Strength calculations of an element compensating circumferential backlash... 255 The equations include the reaction force and friction force components amounting to: RAx = RA sin2ϕ1, RAy = RAcos2ϕ1, RBx = RB sinϕ, RBy = RB cosϕ1, TAx = TAcos2ϕ1, TAy = TA sin2ϕ1, TBx = TB cosϕ1, TBy = TB sinϕ1. Moreover, the action of continuous lo- ad q(ϕ) has been replaced with the concentrated forceQ applied to the point corresponding to angle ϕ = 2/3ϕ0, where: ϕ0 = (2ϕ1 +α) is the maximum angle ϕ value (for the whole beam span). Then the value of forceQ can be calculated from the formula Q= 1 2 q0rϕ0 (3.2) and the force components for the axes x and y amount to Qx = Qsin(ϕ0/3) and Qy =Qcos(ϕ0/3). It appears fromEqs. (3.1) that theanalyzed static system is adoublehyperstatic system.The Menabrei energymethod, according to which the derivative of the system elastic energy relative to the hyperstatic reaction amounts to zero (Zakrzewski and Zawadzki, 1983; Niezgodziński and Niezgodziński, 1996; Dyląg et al., 1999), will be used to determine reactions in the supports and in the beam fixing. The bendingmoment equations and their derivatives over hyperstatic reactionsRA andRB depending on the angle ϕ for particular beam intervals are as follows: — interval I (0<ϕ¬α) MIg(ϕ) =− qr2ϕ2 2ϕ0 sin ϕ 3 ∂MIg ∂RA =0 ∂MIg ∂RB =0 (3.3) — interval II (α<ϕ¬α2 =α+ϕ1) MIIg (ϕ) =− qr2ϕ2 2ϕ0 sin ϕ 3 +RArsin(ϕ−α)−TA[r−rcos(ϕ−α)] ∂MIIg ∂RA = rsin(ϕ−α) ∂MIg ∂RB =0 (3.4) — interval III (α2 <ϕ¬ϕ0) MIIIg (ϕ) =− qr2ϕ2 2ϕ0 sin ϕ 3 +RAr sin(ϕ−α)−TA[r−rcos(ϕ−α)] +RBr sin(ϕ−α2)−TB[r−rcos(ϕ−α2)] ∂MIIIg ∂RA = rsin(ϕ−α) ∂MIIg ∂RB = r sin(ϕ−α2) (3.5) In the above equations, the expression for the bending moment produced by continuous load q(ϕ) = q0ϕ/ϕ0 takes into account equation (3.2): Q(ϕ) = 0.5rϕq(ϕ) = q0rϕ 2/(2ϕ0), as- suming that for any cross section defined by angle ϕ the substitute force Q(ϕ) is applied to the point situated relative to this cross section at angle ϕ/3: MQ(ϕ) = −Q(ϕ)r sin(ϕ/3) = [−qr2ϕ2/(2ϕ0)]sin(ϕ/3). Moreover, in order to simplify the notation, the angle α2 =α+ϕ1 is introduced for determination of the range of variation of angle ϕ in intervals II and III. According to the Menabrei theorem, hyperstatic reactions RA and RB can be calculated from the following system of equations 256 P. Osiński, G. Chruścielski ∂V ∂RA = [ α2 ∫ α ( − qCr 2ϕ2 2ϕ0 sin ϕ 3 +RAr sin(ϕ−α)−TAr[1− cos(ϕ−α) ) [r sin(ϕ−α)] dϕ ] + 1 EI [ ϕC ∫ α2 ( − qCr 2ϕ2 2ϕ0 sin ϕ 3 +RAr sin(ϕ−α)−TAr[1− cos(ϕ−α)] +RBr sin(ϕ−α2)−TBr[1− cos(ϕ−α2)] ) [r sin(ϕ−α)] dϕ ] =0 ∂V ∂RB = 1 EI [ ϕ0 ∫ α2 ( − qCr 2ϕ2 2ϕ0 sin ϕ 3 +RAr sin(ϕ−α)−TAr[1− cos(ϕ−α)] +RBr sin(ϕ−α2)−TBr[1− cos(ϕ−α2)] ) [r sin(ϕ−α2)] dϕ ] =0 (3.6) Having solved the system of equations (3.6), one can calculate reactions RA andRB RA =0.2316∆pbr+0.1915TA −0.0283TB =2376.7N RB =0.3870∆pbr+0.8494TA +0.2997TB =3978.1N (3.7) Then using static equilibrium equations (3.1)one can calculate the reactions in the beam fixing RCx =QCx−RBx−RAx−TC −TBx−TAx =490.4N RCy =QCy −RBy −RAy+TBy +TAy =3628.1N MC =Qr sinϕ03−RBr sinϕ1−RAr sin2ϕ1+TB(r−rcosϕ1) +TA(r−rcos2ϕ1)= 13.35N (3.8) 3.2. Beam with the wrapping angle ϕ0 =132 ◦ 3.2.1. Beam geometry and loading diagram The loading diagram for the beam with the wrapping angle ϕ0 =132 ◦ is shown in Fig. 5a, while its modified version (having pressures replaced with the continuous load) is shown in Fig. 3b. In the latter version, the beam interacts with four teeth of the gear and the working pressure p1 decreases from the initial value p1p = p2 for ϕ=0 up to end value p1k =0.33p2 for ϕ=ϕ0. The continuous load acting on the beam, arising due to the pressure difference, and the resultant substitute force Q is defined the same as for the beamwith the wrapping angle 102◦. Besides the continuous load also the friction forcesTA,TB,TC,TD, reactions in the supports (at the contact with the pump teeth) RA, RB, RC and the reactions in the fixing RDx, RDy, MD (Fig. 5b) act on the beam. 3.2.2. Solution for beam load static system For the loading diagram shown in Fig. 5b, the static equilibrium equations have the form ∑ Px =RDx+RCx+RBx+RAx+TD +TCx+TBx−TAx−Qx =0 ∑ Py =RDy+RCy +RBy −RAy−TCy −TBy −TAy −Qy =0 ∑ MD =MD+RCr sinϕ1+RBr sin2ϕ1+RArcos ϕ1 2 −TC(r−rcosϕ1) −TB(r−rcos2ϕ1)−TA ( r+rsin ϕ1 2 ) −Qrsin ϕ0 3 =0 (3.9) Strength calculations of an element compensating circumferential backlash... 257 Fig. 5. (a) Static diagram of the beamwith the wrapping angle of 132◦, loaded with working pressure p1 and compensating pressure p2, (b) diagram after introduction of continuous load replacing the action of pressures p1 and p2 The equilibrium equations include the reaction and friction force components amoun- ting to: RAx = RAcos(ϕ1/2), RAy = RA sin(ϕ1/2), RBx = RB sin2ϕ1, RBy = RB cos2ϕ1, RCx =RC sinϕ1,RCy =RC cosϕ1,TAx =TA sin(ϕ1/2),TAy =TAcos(ϕ1/2),TBx =TB cos2ϕ1, TBy =TB sin2ϕ1, TCx =TC cosϕ1, TCy =TC sinϕ1,Qx =Qsin(ϕ0/3), Qy =Qcos(ϕ0/3). Since the considered beam is a triple hyperstatic system, in order to calculate the reactions occurring in the supports and in the beam fixing one should formulate three equations based on the Menabrei method. Assuming RA, RB and RC as hyperstatic reactions occurring in the supports, the equations become ∂V ∂RA = 1 EI [ α2 ∫ α ( − q0r 2ϕ2 2ϕ0 sin ϕ 3 +RAr sin(ϕ−α)−TAr[1− cos(ϕ−α)] ) [r sin(ϕ−α)] dϕ ] + 1 EI [ α3 ∫ α2 ( − q0r 2ϕ2 2ϕ0 sin ϕ 3 +RArsin(ϕ−α)−TAr[1− cos(ϕ−α)] +RBrsin(ϕ−α2)−TBr[1− cos(ϕ−α2)] ) [r sin(ϕ−α)] dϕ ] + 1 EI [ ϕ0 ∫ α3 ( − q0r 2ϕ2 2ϕC sin ϕ 3 +RArsin(ϕ−α)−TAr[1− cos(ϕ−α)]+RBr sin(ϕ−α2) −TBr[1− cos(ϕ−α2)]+RCr sin(ϕ−α3)−TCr[1− cos(ϕ−α3)] ) [r sin(ϕ−α)] dϕ ] =0 ∂V ∂RB = 1 EI [ α3 ∫ α2 ( − q0r 2ϕ2 2ϕ0 sin ϕ 3 +RArsin(ϕ−α)−TAr[1− cos(ϕ−α)] (3.10) +RBrsin(ϕ−α2)−TBr[1− cos(ϕ−α2)] ) [r sin(ϕ−α2)] dϕ ] + 1 EI [ ϕ0 ∫ α3 ( − q0r 2ϕ2 2ϕC sin ϕ 3 +RArsin(ϕ−α)−TAr[1− cos(ϕ−α)]+RBr sin(ϕ−α2) −TBr[1− cos(ϕ−α2)]+RCr sin(ϕ−α3)−TCr[1− cos(ϕ−α3)] ) [r sin(ϕ−α2)] dϕ ] =0 258 P. Osiński, G. Chruścielski ∂V ∂RC = 1 EI [ ϕ0 ∫ α3 ( − q0r 2ϕ2 2ϕ0 sin ϕ 3 +RArsin(ϕ−α)−TAr[1− cos(ϕ−α)]+RBr sin(ϕ−α2) −TBr[1− cos(ϕ−α2)]+RCr sin(ϕ−α3)−TCr[1− cos(ϕ−α3)] ) [r sin(ϕ−α3)] dϕ ] =0 Having solved equations (3.10), one can calculate reactions RA,RB andRC RA =0.1103∆pbr+0.2522TA −0.0259TB +0.0067TC =1510.2N RB =0.3164∆pbr+0.7156TA −0.3009TB −0.0369TC =4334.5N RC =0.5946∆pbr+0.6019TA−0.6787TB +0.3009TC =8144.1N (3.11) Finally, using static equilibrium equations (3.9), one gets the values of reactions in the beam fixing RDx =Qx−RCx−RBx−RAx−TD−TCx−TBx+TAx =583.8N RDy =Qy−RCy −RBy +RAy+TCy +TBy +TAy =3889.9N MD =Qrsin ϕ0 3 −RCr sinϕ1−RBr sin2ϕ1−RArcos ϕ1 2 +TC(r−rcosϕ1) +TB(r−rcos2ϕ1)+TA ( r+rsin ϕ1 2 ) =15.98N (3.12) 3.3. Beam with the wrapping angle ϕ0 =169 ◦ 3.3.1. Beam geometry and loading diagram Figures 6a and 6b show the loading diagram and the diagram which takes into account the replacement of pressures p1 and p2 (acting on both sides of the beam) with continuous load q for the pumpwith the compensating chamber with the wrapping angle ϕ0 =169 ◦. Fig. 6. (a) Static diagram of the beamwith the wrapping angleϕ0 =169 ◦, loaded with working pressure p1 and compensating pressure p2, (b) diagram after introduction of continuous load replacing the action of pressures p1 and p2 In this case, the chamber beam is in contact with five teeth of the gear and the wor- king pressure p1 decreases from the initial value p1p = p2 for ϕ = 0 down to the end value p1k = 0.166p2 for ϕ=ϕ0. Besides the continuous load, as shown in Fig. 4b, the friction forces TA,TB,TC, TD,TE, reaction forces in the supportsRA,RB,RC,RD and the forces in the beam fixingREx,REy,ME act on the beam. Strength calculations of an element compensating circumferential backlash... 259 3.3.2. Solutions for the beam subject to load static scheme The static equilibrium equations for the considered beam assume the form ∑ Px =REx+RDx+RCx+RBx+RAx+TE +TDx+TCx−TBx−TAx−Qx =0 ∑ Py =REy+RDy+RCy −RBy −RAy−TDy −TCy−TBy −TAy −Qy =0 ∑ ME =ME +RDrsinϕ1+RCr sin2ϕ1+RBrcos ϕ1 2 +RArcos 3ϕ1 2 −TD(r−rcosϕ1)−TC(r−rcos2ϕ1)−TB ( r+r sin ϕ1 2 ) −TA ( r+rsin 3ϕ1 2 ) −Qr sin ϕc 3 =0 (3.13) The reaction and friction force components in equations (3.13) are described by the formu- las: RAx = RAcos(3ϕ1/2), RAy = RA sin(3ϕ1/2), RBx = RB cos(ϕ1/2), RBy = RB sin(ϕ1/2), RCx =RC sin2ϕ1,RCy =RC cos2ϕ1,RDx =RD sinϕ1,RDy =RD cosϕ1,TAx =TA sin(3ϕ1/2), TAy = TAcos(3ϕ1/2), TBx = TB sin(ϕ1/2), TBy = TB cos(ϕ1/2), TCx = TC cos2ϕ1, TCy =TC sin2ϕ1, TDx =TD cosϕ1, TDy =TD sinϕ1,Qx =Qsin(ϕ0/3), Qy =Qcos(ϕ0/3). In order to determine the four hyperstatic reactions (assumed here as the reactions in the supports) one should formulate fourMenabrei equations ∂V ∂RA = 1 EI [ α2 ∫ α ( − q0r 2ϕ2 2ϕ0 sin ϕ 3 +RArsin(ϕ−α)−TAr[1− cos(ϕ−α)] ) [r sin(ϕ−α)] dϕ ] + 1 EI [ α3 ∫ α2 − q0r 2ϕ2 2ϕ0 sin ϕ 3 +RAr sin(ϕ−α)−TAr[1− cos(ϕ−α)] +RBr sin(ϕ−α2)−TBr[1− cos(ϕ−α2)] ) [r sin(ϕ−α)] dϕ ] + 1 EI [ α4 ∫ α3 ( − q0r 2ϕ2 2ϕ0 sin ϕ 3 +RArsin(ϕ−α)−TAr[1− cos(ϕ−α)]+RBr sin(ϕ−α2) −TBr[1− cos(ϕ−α2)]+RCrsin(ϕ−α3)−TCr[1− cos(ϕ−α3)] ) [r sin(ϕ−α)] dϕ ] + 1 EI [ ϕ0 ∫ α4 ( − q0r 2ϕ2 2ϕ0 sin ϕ 3 +RArsin(ϕ−α)−TAr[1− cos(ϕ−α)] +RBr sin(ϕ−α2)−TBr[1− cos(ϕ−α2)]+RCrsin(ϕ−α3)−TCr[1− cos(ϕ−α3)] +RDr sin(ϕ−α4)−TDr[1− cos(ϕ−α4)] ) [r sin(ϕ−α)] dϕ ] =0 ∂V ∂RB = 1 EI [ α3 ∫ α2 ( − q0r 2ϕ2 2ϕ0 sin ϕ 3 +RAr sin(ϕ−α)−TAr[1− cos(ϕ−α)] +RBr sin(ϕ−α2)−TBr[1− cos(ϕ−α2)] ) [r sin(ϕ−α2)] dϕ ] + 1 EI [ α4 ∫ α3 ( − q0r 2ϕ2 2ϕ0 sin ϕ 3 +RArsin(ϕ−α)−TAr[1− cos(ϕ−α)]+RBr sin(ϕ−α2) 260 P. Osiński, G. Chruścielski −TBr[1− cos(ϕ−α2)]+RCrsin(ϕ−α3)−TCr[1− cos(ϕ−α3)] ) [r sin(ϕ−α2)] dϕ ] + 1 EI [ ϕ0 ∫ α4 ( − q0r 2ϕ2 2ϕ0 sin ϕ 3 +RArsin(ϕ−α)−TAr[1− cos(ϕ−α)] (3.14) +RBrsin(ϕ−α2)−TBr[1− cos(ϕ−α2)]+RCrsin(ϕ−α3)−TCr[1− cos(ϕ−α3)] +RDr sin(ϕ−α4)−TCr[1− cos(ϕ−α4)] ) [r sin(ϕ−α2)] dϕ ] =0 ∂V ∂RC = 1 EI [ α4 ∫ α3 ( − q0r 2ϕ2 2ϕ0 sin ϕ 3 +RAr sin(ϕ−α)−TAr[1− cos(ϕ−α)] +RBrsin(ϕ−α2)−TBr[1− cos(ϕ−α2)]+RCrsin(ϕ−α3) −TCr[1− cos(ϕ−α3)] ) [r sin(ϕ−α3)] dϕ ] + 1 EI [ ϕ0 ∫ α4 ( − q0r 2ϕ2 2ϕ0 sin ϕ 3 +RArsin(ϕ−α)−TAr[1− cos(ϕ−α)] +RBrsin(ϕ−α2)−TBr[1− cos(ϕ−α2)]+RCrsin(ϕ−α3)−TCr[1− cos(ϕ−α3)] +RDr sin(ϕ−α4)−TDr[1− cos(ϕ−α4)] ) [r sin(ϕ−α3)] dϕ ] =0 ∂V ∂RD = 1 EI [ ϕ0 ∫ α4 ( − q0r 2ϕ2 2ϕ0 sin ϕ 3 +RAr sin(ϕ−α)−TAr[1− cos(ϕ−α)] +RBrsin(ϕ−α2)−TBr[1− cos(ϕ−α2)]+RCrsin(ϕ−α3)−TCr[1− cos(ϕ−α3)] +RDr sin(ϕ−α4)−TCDr[1− cos(ϕ−α4)] ) [r sin(ϕ−α4)] dϕ ] =0 By solving theMenabrei equations one can determine reactions RA,RB,RC andRD RA =0.0908∆pbr−0.4034TA −0.0255TB +0.0073TC −0.0019TD =1548.2N RB =0.2542∆pbr+0.7128TA +0.3001TB−0.0400TC +0.0104TD =4352.4N RC =0.4593∆pbr+0.6065TA+0.6777TB +0.3127TC −0.0400TD =7863.2N RD =0.7524∆pbr+0.6340TA +0.6126TB +0.6750TC +0.3027TD =12880.2N (3.15) Finally, using static equilibrium equations (3.13), one can calculate reactions in the beam fixing REx =Qx−RDx−RCx−RBx−RAx−TE −TDx−TCx−TBx−TAx =877.0N REy =Qy−RDy−RCy +RBy +RAy+TDy +TCy +TBy +TAy =3747.4N ME =Qr sin ϕ0 3 −RDrsinϕ1−RCr sin2ϕ1−RBrcos ϕ1 2 −RArcos 3ϕ1 2 +TD(r−rcosϕ1)+TC(r−rcos2ϕ1)+TB ( r+r sin ϕ1 2 ) +TA ( r+rsin 3ϕ1 2 ) =23.8Nm (3.16) Strength calculations of an element compensating circumferential backlash... 261 4. Calculations of the minimal beam thickness in a fixed cross section The beam thickness hmust satisfy the strength condition for the fixed beam cross section. This cross section is loaded with reaction forces RC, RD or RE for the pump geometric conditions atϕ0 respectively 102 ◦, 132◦, 169◦ and the bendingmomentMC,MD orME, respectively. The loads generate a complex state of stress in the fixed cross section. The shearing stress (produced by the tangential component of the reaction denoted generally asRy) reaches the highest value in central fibres of the cross sectionwhile the bending stress reaches the highest value in extreme fibres. The strength condition concerning the maximum shear stress has the form τmax = 3 2 |Ry| bh ¬ kt (4.1) thus the minimal beam thickness must satisfy the criterion hmin(τ) = 3 2 |Ry| bkt (4.2) The strength condition for the allowable normal stress must take into account the simulta- neous action of the tensile (or compressive) stress produced by the reaction componentRx and the bending stress generated by the fixingmomentMu in the extreme fibres σ= |σr|+ |σg| ¬ kr (4.3) By substituting thenormal stress values (calculated as for straight barswhenbeamcurvature radius r > 6h, the error due to the shift of the neutral beambending axis does not exceed 0.5%) σr = Rx/(bh), σg = 6Mu/(bh 2) into equation (4.3), one gets the following equation for the minimal beam thickness hmin bkrh 2 min−|Rx|hmin−6|Mu|=0 (4.4) The solution of this equation yields the second value of the minimal beam thickness hmin(τ) = |Rx|+ √ R2x+24Mubkr 2bkr (4.5) One should adopt the second value of the two values obtained from formulas (4.2) and (4.5) as the minimal beamwidth h. Table 1 shows exemplary minimal beam thickness values calculated for three geometric ver- sions of the pump, assuming the experimental friction forces generated by teeth of the gears Ti =7.96N, safety factor n=1.4 and permissible stresses: kr =350MPa and kt =400MPa. Table 1.Exemplary load values and theminimal thicknesshmin for thebeamwith thewrapping angle ϕ0 =102 ◦ Beam wrapping r b p1k p2 hmin angle ϕ0 =102 ◦ 26mm 26.3mm 15MPa 30MPa 2.98mm ϕ0 =132 ◦ 26mm 26.3mm 10MPa 30MPa 3.26mm ϕ0 =169 ◦ 26mm 26.3mm 5MPa 30MPa 3.99mm The results of the static strength calculations show that from among the three versions of the compensating chamber the most advantageous one is the version with the beam with thw wrapping angle ϕ0 = 102 ◦ for which the minimal beam thickness amounts to 2.98mm. In the case of the other versions, the beam thickness needs to be increased: 262 P. Osiński, G. Chruścielski • for the beam with the wrapping angle ϕ0 = 132 ◦, the minimal thickness amounts to 3.26mm (an increase by 9.4%), • for the beam with the wrapping angle ϕ0 = 169 ◦, the minimal thickness amounts to 3.99mm (an increase by 31%). Experimental studies are planned to be carried out on prototypes of the pump in order to verify the results of the calculations. References 1. Chrobot M., Łabik M., Siudy B., Wieczorek K., 1997, Hydraulic optimization of axial backlash compensation (in Polish),Report Series Report, No. S-019/97 2. Dyląg Z., Jakubowicz A., Orłoś Z., 1999, Strength of Materials (in Polish), Vol. I/II,WNT, Warsaw 3. Judin E.M., 1958,Gear Pumps (in Polish), PWT,Warsaw 4. Kollek W., 2004,Fundamental of the Design of Hydraulic Drives and Controls (in Polish),Wro- cławUniversity of Technology Publishing House,Wrocław 5. Kollek W., 1996,Gear Pumps – Design and Operation (in Polish), Ossolineum,Wrocław 6. Kollek W., Radziwanowska U., 2015, Energetic efficiency of gear micropumps, Archives of Civil and Mechanical Engineering, 15, 1, 109-115 7. NiezgodzińskiM.E.,NiezgodzińskiT., 1996,Strength ofMaterials (inPolish),PWN,Warsaw 8. Osiński P., 2012a, An external gear pump (in Polish), Patent application No. P398436,Warsaw 9. Osiński P., 2012b, An external gear pump (in Polish), Patent application No. P398437,Warsaw 10. Osiński P., 2013,High-Pressure and Low-Fluctuation Internal Gear Pumps (in Polish) Wrocław University of Technology Publishing House,Wrocław 11. Osiński P., Deptuła A., Partyka M.A., 2013, Discrete optimization of a gear pump after tooth root undercutting by means of multi-valued logic trees, Archives of Civil and Mechanical Engineering, 13, 4, 422-431, http://dx.doi.org/10.1016/j.acme.2013.05.001 12. Osiński P., Kania A., Kollek W., Wiczkowski E., 2012, An external gear pump (in Polish), Patent application No. P397539,Warsaw 13. Osiński P., Kollek W., 2007, Modelling and Design of Gear Pumps, Wrocław University of Technology Publishing House,Wrocław 14. Ragunathan C., Manoharan C., 2012, Dynamic analysis of hydrodynamic gear pump perfor- mance using design of experiment stand operational parameters, IOSR Journal of Mechanical and Civil Engineering, 1, 6, 17–23 15. Singal R.K., Singal M., Singal R., 2009,Hydraulic Machines. Fluid Machinery, I.K. Interna- tional Publishing House, NewDelhi 16. Stryczek S., 1995,Hydrostatic Drives (in Polish), Vol. I, II. WNT, Warsaw 17. Vacca A., Guidetti M., 2011, Modelling and experimental validation of external spur gear machines for fluid power applications, Simulation Modelling Practice and Theory, 19, 2007-2031 18. Wang S., Sakurai H., Kasarekar A., 2011, The optimal design in external gear pumps and motors,ASME Transactions on Mechatronics, 16, 5. 19. Wiczkowski E., 2012, An external gear pump (in Polish), Patent application No. P397540, Warsaw 20. Zakrzewski M., Zawadzki J., 1983, Strength of Materials (in Polish), PWN,Warsaw Manuscript received December 7, 2014; accepted for print August 13, 2015