Page 433 - Cam Design Handbook
P. 433

THB13  9/19/03  7:56 PM  Page 421

                                CAM SYSTEM DYNAMICS—RESPONSE               421

            13.5.7 Introduction to System Optimization

            Having formulated an optimization criterion, we are now faced with its evaluation. Various
            linear techniques have been used in the design optimization of high-speed cam-follower
            systems.
               Linear methods are certainly preferable when applicable. However, for the synthesis
            of cam-follower systems with nonlinearities such as contact stresses, pressure angles, cam
            curvature, and follower guide friction, more general methods are needed. In this section
            the nonlinearities are handled by means of optimal-control theory. Although this method
            is powerful, it requires a well-defined problem. In certain instances an optimal solution
            does  not  exist.  The  solution  of  an  optimization  problem  using  optimal  control  theory
            always results in a two-point boundary-value problem. Such problems can be tedious and
            time consuming. Except for simple linear problems, the vast majority require a numerical
            solution. Such a procedure will be described.



            13.5.8 Optimization of Tuned Dwell-Rise-Dwell Cams
            13.5.8.1 Formulation  of  Two-Point  Boundary-Value  Problems.  The  optimality
            criteria for high-speed cam-follower systems derived previously will now be used for the
            synthesis of a tuned DRD cam. The optimization problem can be stated as the optimality
            criterion (Eq. [13.42]) subject to the system equation (Eq. [13.30]) and boundary condi-
            tions (Eq. [13.32]).
               By defining new state variables:
                                                 Y
                                            x
                                            È ˘  È ˘
                                             1
                                            Í ˙  Í ˙
                                                  ˙
                                            x
                                                 Y
                                            Í ˙  Í ˙
                                             2
                                                 Í ˙
                                       x t () = Í ˙ = Y c                (13.43)
                                            x
                                             3
                                            Í ˙  Í ˙
                                            x
                                                 Y
                                            Í ˙  Í ˙ ˙ c
                                             4
                                            Í ˙  Í ˙
                                                  ˙˙
                                            x
                                            Î ˚
                                                 Î Í ˚ ˙
                                                 Y
                                             5
                                                  c
                                  .
            and the control variable, u = x 5 , the parameters in Eq. (13.42) can be expressed as follows.
                             F =- x +  K x                  ¸
                              fe   1   k 3
                                                            Ô
                              ˙˙
                             F =  r x +  r x -  r x +  K x  ˝            (13.44)
                              fe  0 1  1 2  0 3  k  5
                              ˙˙˙
                                               r x +
                                    -
                             F =  K u r rx -( r - )  r rx -  r x  Ô
                                            2
                                                            ˛
                              fe   k  011   1  0  2  0 1 3  0 4
            where
                                        r = ( 2pl ) 2
                                         0
                                           22 (
                                        r = zpl  ) 2
                                         1
            and
                                               )
                                       ( ,, ,
                                                  ( , ,
                                 N =  S x x xx C x xx ).                 (13.45)
                                  F   F  1  3  4  5  F  3  4  5
            The system equation and boundary conditions can also be written as
   428   429   430   431   432   433   434   435   436   437   438