Page 221 - The Combined Finite-Discrete Element Method
P. 221

204    TEMPORAL DISCRETISATION

            • The D-1/12 Time Integration Scheme (D-1/12): The recursive formula for D-1/12 is
              as follows:

                                                    f t  2
                                    x t − x t−h + v t h +  h
                                                    2 m               1 df
                                                                ˜
                       x t+h = x t−h +                    where b t =           (5.106)
                                             1                       m dx
                                               ˜
                                         1 −   b t  t 2                     t
                                             12
                                    h
                       v t+h = v t +  (8f t + 5f t+h − f t−h )                  (5.107)
                                  12 m
            5.3.2  Gear’s predictor-corrector time integration schemes (PC-3,
                   PC-4, and PC-5)

            The third order scheme is referred to as PC-3, while the fourth and fifth order schemes
            are referred to as PC-4 and PC-5, respectively. All three schemes use the same recursive
            formula, which comprises three stages:
            (a) Prediction: in the prediction stage, positions are calculated at time (t + ∆t) by means
               of a Taylor series based on positions and their derivatives:
                                                 h 2  ... h 3  iv  t 4  v  h 5
                              x t+h,p = x t +˙x t h +¨x t  + x t  + x t  + x t  (5.108)
                                                 2!     3!     4!     5!
                                               ... h 2  iv  h 3  v  h 4
                              ˙ x t+h,p =˙x t +¨x t h + x t  + x t  + x t
                                                 2!      3!    4!
                                          ...   iv x 2  v  x 3
                              ¨ x t+h,p =¨x t + x t h + x  + x
                                                t       t
                                                  2!     3!
                              ...     ...  iv    v  h 2
                              x t+h,p = x t + x h + x t
                                           t
                                                  2!
                                            v
                                       iv
                               iv
                              x t+h,p  = x + x h
                                       t
                                            t
                             x v t+ t,p  = x t v
            (b) Evaluation: in the evaluation stage, the force f t+h at time (t + h) is evaluated using
               the x t+h,p position from the prediction stage. From this force, acceleration ¨x t+h at time
               (t + h) is calculated, and the discrepancy between this acceleration and the predicted
               acceleration ¨x t+h,p is evaluated, i.e.
                                          ¨x = (¨x t+h −¨x t+h,p )              (5.109)


            (c) Correction: in the correction stage, the predicted positions and time derivatives are
               corrected using the following formulae:

                                                         ¨xh 2
                                        x t+h = x t+h,p + α 0                   (5.110)
                                                          2!
                                                          ¨xh 2
                                        ˙ x t+h h =˙x t+h,p h + α 1
                                                           2!
   216   217   218   219   220   221   222   223   224   225   226