液压支架的优化设计英文翻译.pdf
Struct Multidisc Optim 20, 7682 Springer-Verlag 2000 Optimal design of hydraulic support M. Oblak, B. Harl and B. Butinar Abstract This paper describes a procedure for optimal determination of two groups of parameters of a hydraulic support employed in the mining industry. The procedure is based on mathematical programming methods. In the fi rst step, the optimal values of some parameters of the leading four-bar mechanism are found in order to ensure the desired motion of the support with minimal transver- sal displacements. In the second step, maximal tolerances of the optimal values of the leading four-bar mechanism are calculated, so the response of hydraulic support will be satisfying. Key words four-bar mechanism, optimal design, math- ematical programming,approximationmethod, tolerance 1 Introduction The designer aims to fi nd the best design for the mechan- ical system considered. Part of this eff ort is the optimal choice of some selected parameters of a system. Methods of mathematical programming can be used, if a suitable mathematical model of the system is made. Of course, it depends on the type of the system. With this formulation, good computer support is assured to look for optimal pa- rameters of the system. The hydraulic support (Fig. 1) described by Harl (1998) is a part of the mining industry equipment in the mine Velenje-Slovenia, used for protection of work- ing environment in the gallery. It consists of two four-bar Received April 13, 1999 M. Oblak1, B. Harl2and B. Butinar3 1 Faculty of Mechanical Engineering, Smetanova 17, 2000 Maribor, Slovenia e-mail: maks.oblakuni-mb.si 2 M.P.P. Razvoj d.o.o., Ptujska 184, 2000 Maribor, Slovenia e-mail: bostjan.harluni-mb.si 3 Faculty of Chemistry and Chemical Engineering, Smetanova 17, 2000 Maribor, Slovenia e-mail: branko.butinaruni-mb.si mechanisms FEDG and AEDB as shown in Fig. 2. The mechanism AEDB defi nes the path of coupler point C and the mechanism FEDG is used to drive the support by a hydraulic actuator. Fig. 1 Hydraulic support It is required that the motion of the support, more precisely, the motion of point C in Fig. 2, is vertical with minimal transversal displacements. If this is not the case, the hydraulic support will not work properly because it is stranded on removal of the earth machine. A prototype of the hydraulic support was tested in a laboratory (Grm 1992). The support exhibited large transversal displacements, which would reduce its em- ployability. Therefore, a redesign was necessary. The project should be improved with minimal cost if pos- 77 Fig. 2 Two four-bar mechanisms sible. It was decided to fi nd the best values for the most problematic parameters a1,a2,a4of the leading four-bar mechanism AEDB with methods of mathematical pro- gramming. Otherwise it would be necessary to change the project, at least mechanism AEDB. The solution of above problem will give us the re- sponse of hydraulic support for the ideal system. Real response will be diff erent because of tolerances of vari- ous parameters of the system, which is why the maximal allowed tolerances of parameters a1,a2,a4will be calcu- lated, with help of methods of mathematical program- ming. 2 The deterministic model of the hydraulic support At fi rst it is necessary to develop an appropriate mechan- ical model of the hydraulic support. It could be based on the following assumptions: the links are rigid bodies, the motion of individual links is relatively slow. The hydraulic support is a mechanism with one de- gree of freedom. Its kinematics can be modelled with syn- chronous motion of two four-bar mechanisms FEDG and AEDB (Oblak et al. 1998). The leading four-bar mech- anism AEDB has a decisive infl uence on the motion of the hydraulic support. Mechanism 2 is used to drive the support by a hydraulic actuator. The motion of the sup- port is well described by the trajectory L of the coupler point C. Therefore, the task is to fi nd the optimal values of link lengths of mechanism 1 by requiring that the tra- jectory of the point C is as near as possible to the desired trajectory K. The synthesis of the four-bar mechanism 1 has been performed with help of kinematics equations of motion givenby Rao and Dukkipati (1989).The generalsituation is depicted in Fig. 3. Fig. 3 Trajectory L of the point C Equations of trajectory L of the point C will be writ- ten in the coordinate frame considered. Coordinates x and y of the point C will be written with the typical parameters of a four-bar mechanism a1,a2, ., a6. The coordinates of points B and D are xB= xa5cos,(1) yB= ya5sin,(2) xD= xa6cos(+),(3) yD= ya6sin(+).(4) The parameters a1,a2, ., a6are related to each other by x2 B+y 2 B= a 2 2, (5) (xDa1)2+y2 D= a 2 4. (6) By substituting (1)(4) into (5)(6) the response equations of the support are obtained as (xa5cos)2+(ya5sin)2a2 2= 0, (7) xa6cos(+)a12+ ya6sin(+)2a2 4= 0. (8) This equation representsthe base of the mathematical model forcalculatingthe optimalvalues ofparametersa1, a2, a4. 78 2.1 Mathematical model The mathematical model ofthe systemwill be formulated in the form proposed by Haug and Arora (1979): min f(u,v),(9) subject to constraints gi(u,v) 0,i = 1,2,. ,?,(10) and response equations hj(u,v) = 0,j = 1,2,. ,m.(11) The vector u = u1.unTis called the vector of design variables, v = v1.vmTis the vector of response vari- ables and f in (9) is the objective function. To perform the optimal design of the leading four-bar mechanism AEDB, the vector of design variables is de- fi ned as u = a1a2a4T,(12) and the vector of response variables as v = x yT.(13) The dimensions a3, a5, a6of the corresponding links are kept fi xed. The objective function is defi ned as some “measure of diff erence” between the trajectory L and the desired tra- jectory K as f(u,v) = maxg0(y)f0(y)2,(14) where x = g0(y) is the equation of the curve K and x = f0(y) is the equation of the curve L. Suitable limitations for our system will be chosen. The system must satisfy the well-known Grasshoff conditions (a3+a4)(a1+a2) 0,(15) (a2+a3)(a1+a4) 0.(16) Inequalities (15) and (16) express the property of a four- bar mechanism, where the links a2,a4may only oscillate. The condition u u u(17) prescribes the lower and upper bounds of the design vari- ables. The problem (9)(11) is not directly solvable with the usual gradient-based optimization methods. This could be circumvented by introducing an artifi cial design vari- able un+1as proposed by Hsieh and Arora (1984). The new formulation exhibiting a more convenient form may be written as min un+1,(18) subject to gi(u,v) 0,i = 1,2,. ,?,(19) f(u,v)un+1 0,(20) and response equations hj(u,v) = 0,j = 1,2,. ,m,(21) where u = u1.unun+1Tand v = v1.vmT. Anonlinearprogrammingproblemofthe leading four- bar mechanism AEDB can therefore be defi ned as min a7,(22) subject to constraints (a3+a4)(a1+a2) 0,(23) (a2+a3)(a1+a4) 0,(24) a1 a1 a1,a2 a2 a2, a4 a4 a4,(25) g0(y)f0(y)2a7 0,(y ? ?y,y?), (26) and response equations (xa5cos)2+(ya5sin)2a2 2= 0, (27) xa6cos(+)a12+ ya6sin(+)2a2 4= 0. (28) This formulation enables the minimization of the diff er- ence between the transversal displacement of the point C and the prescribed trajectory K. The result is the optimal values of the parameters a1, a2, a4. 79 3 The stochastic model of the hydraulic support The mathematical model (22)(28) may be used to cal- culate such values of the parameters a1, a2, a4, that the “diff erence between trajectories L and K” is mini- mal. However, the real trajectory L of the point C could deviate from the calculated values because of diff erent infl uences. The suitable mathematical model deviation could be treated dependently on tolerances of parameters a1,a2,a4. The response equations (27)(28) allow us to calcu- late the vector of response variables v in dependence on the vector of design variables u. This implies v =h(u). The function h is the base of the mathematical model (22)(28), because it represents the relationship between the vector of design variables u and response v of our mechanical system. The same function h can be used to calculate the maximal allowed values of the tolerances a1, a2, a4of parameters a1, a2, a4. In the stochastic model the vector u = u1.unTof design variables is treated as a random vector U = U1 .UnT, meaning that the vector v = v1.vmTof re- sponse variablesis alsoa randomvectorV = V1.VmT, V =h(U).(29) It is supposed that the design variables U1, ., Unare independent from the probability point of view and that they exhibit normal distribution, Uk N(µk,k) (k = 1,2,. ,n). The main parameters µkand k(k = 1, 2, . , n) could be bound with technological notions such as nominal measures, µk= ukand tolerances, e.g. uk= 3k, so events µkuk Uk µk+uk,k = 1,2,. ,n,(30) will occur with the chosen probability. The probability distribution function of the random vector V, that is searched for depends on the probabil- ity distribution function of the random vector U and it is practically impossible to calculate. Therefore, the ran- dom vector V will be described with help of “numbers characteristics”, that can be estimated by Taylor approx- imation of the function h in the point u = u1.unTor with help of the Monte Carlo method in the papers by Oblak (1982) and Harl (1998). 3.1 The mathematical model The mathematical model for calculating optimal toler- ances of the hydraulic support will be formulated as a nonlinear programmingproblemwith independent vari- ables w = a1a2a4T,(31) and objective function f(w) = 1 a1 + 1 a2 + 1 a4 (32) with conditions YE 0,(33) a1 a1 a1,a2 a2 a2, a4 a4 a4.(34) In (33) E is the maximal allowed standard deviation Y of coordinate x of the point C and Y= 1 6 ? ? ? ? jA ? g1 aj (µ1,µ2,µ4) ?2 aj, A = 1,2,4.(35) The nonlinear programming problem for calculating the optimal tolerances could be therefore defi ned as min ? 1 a1 + 1 a2 + 1 a4 ? ,(36) subject to constraints YE 0,(37) a1 a1 a1,a2 a2 a2, a4 a4 a4.(38) 4 Numerical example The carrying capability of the hydraulic support is 1600kN. Both four-bar mechanisms AEDB and FEDG must fulfi ll the following demand: they must allow minimal transversal displacements of the point C, and they must provide suffi cient side stability. The parameters of the hydraulic support (Fig. 2) are given in Table 1. The drive mechanism FEDG is specifi ed by the vector b1,b2,b3,b4T= 400,(1325+d),1251,1310T(mm), (39) and the mechanism AEDB by a1,a2,a3,a4T= 674,1360,382,1310T(mm).(40) In (39), the parameter d is a walk of the support with maximal value of 925mm. Parameters for the shaft of the mechanism AEDB are given in Table 2. 80 Table 1 Parameters of hydraulic support SignLength (mm) M110 N510 O640 P430 Q200 S1415 T380 Table 2 Parameters of the shaft for mechanism AEDB Sign a51427.70 mm a61809.68 mm 179.34 0.52 0.14 4.1 Optimal links of mechanism AEDB With this data the mathematical model of the four-bar mechanisms AEDB could be written in the form of (22) (28). A straight line is defi ned by x = 65 (mm) (Fig. 3) for the desired trajectory of the point C. That is why condi- tion (26) is (x65)a7 0.(41) The angle between links AB and AE may vary be- tween 76.8and 94.8. The condition (41) will be dis- cretized by taking into accountonly the points x1, x2, ., x19in Table 3. These points correspond to the angles 21, 22, ., 219ofthe interval 76.8, 94.8 at regularinter- vals of 1. The lower and upper bounds of design variables are u = 640,1330,1280,0T(mm),(42) u = 700,1390,1340,30T(mm).(43) The nonlinear programming problem is formulated in the form of (22)(28). The problem is solved by the op- timizer described by Kegl et al. (1991) based on approx- imation method. The design derivatives are calculated numerically by using the direct diff erentiation method. The starting values of design variables are ?0a 1,0a2,0a4,0a7 ?T = 674,1360,1310,30T(mm).(44) The optimal design parameters after 25 iterations are u= 676.42,1360.74,1309.88,3.65T(mm).(45) In Table 3 the coordinates x and y of the coupler point C are listed for the starting and optimal designs, respec- tively. Table 3 Coordinates x and y of the point C Anglexstartystartxendyend 2()(mm)(mm)(mm)(mm) 76.866.781784.8769.471787.50 77.865.911817.6768.741820.40 78.864.951850.0967.931852.92 79.863.921882.1567.041885.07 80.862.841913.8566.121916.87 81.861.751945.2065.201948.32 82.860.671976.2264.291979.44 83.859.652006.9163.462010.23 84.858.722037.2862.722040.70 85.857.922067.3562.132070.87 86.857.302097.1161.732100.74 87.856.912126.5961.572130.32 88.856.812155.8061.722159.63 89.857.062184.7462.242188.67 90.857.732213.4263.212217.46 91.858.912241.8764.712246.01 92.860.712270.0866.852274.33 93.863.212298.0969.732302.44 94.866.562325.8970.502330.36 Figure4illustrates the trajectoriesLofthe pointC for the starting (hatched) and optimal (full) design as well as the straight line K. 4.2 Optimal tolerances for mechanism AEDB In the nonlinear programming problem (36)(38), the chosen lower and upper bounds of independent variables a1, a2, a4are w = 0.001,0.001,0.001T(mm),(46) w = 3.0,3.0,3.0T(mm).(47) The starting values of the independent variables are w0= 0.1,0.1,0.1T(mm).(48) The allowed deviation of the trajectory was chosen for two cases as E = 0.01 and E = 0.05. In the fi rst case, the 81 Fig. 4 Trajectories of the point C Table 4 Optimal tolerances for E = 0.01 SignValue (mm) a10.01917 a20.00868 a40.00933 Table 5 Optimal tolerances for E = 0.05 SignValue (mm) a10.09855 a20.04339 a40.04667 Fig. 5 Standard deviations for E = 0.01 optimal tolerances for the design variables a1, a2, a4were calculated after 9 iterations. For E = 0.05 the optimum was obtained after 7 iterations. The results are given in Tables 4 and 5. In Figs. 5 and 6 the standard deviations are calculated by the Monte Carlo method and with Taylor approxima- tion (full line represented Taylor approximation), respec- tively. Fig. 6 Standard deviations for E = 0.05 5 Conclusions With a suitable mathematical model ofthe systemand by employing mathematical programming, the design of the 82 hydraulic support was improved, and better performance was achieved. However, due to the results of optimal tol- erances, it might be reasonable to take into consideration a new construction. This is especially true for the mech- anism AEDB, since very small tolerances raise the costs of production. References Grm, V. 1992: Optimal synthesis of four-bar mechanism. MSc. Thesis. Faculty of Mechanical Engineering Maribor Harl, B. 1998: Stochastic analyses of hydraulic support 2S. MSc. Thesis. Faculty of Mechanical Engineering Maribor Haug, E.J.; Arora, J.S. 1979: Applied optimal design. New York: Wiley Hsieh, C.; Arora, J. 1984: Design sensitivity analysis and op- timisation of dynamic response. Comp
