Page 330 - Advances in Biomechanics and Tissue Regeneration
P. 330

328                           16. ON THE SIMULATION OF ORGAN-ON-CHIP CELL PROCESSES

              Let us define

                                                     f i ¼ p i  vC i , i ¼ 1,…,n
                                                   f n + i ¼ q i  vS i , i ¼ 1,…,m                         (16.86)
                                       T
           and f ¼ (f 1 , …, f n , f n+1 , …, f n+m ) .
              Finally let us define

                                                            n        n
                                                 ∂v        X        X
                                        s i  ¼ C i  + C i F i    C i F ij +  C j F ji , i ¼ 1,…,n
                                                 ∂x
                                                          j ¼ 1     j ¼ 1
                                                          j 6¼ i    j 6¼ i
                                                     n        n
                                                ∂v  X        X
                                       s n + i ¼ S i     C j G ij +  C j G ji , i ¼ 1,…,m                  (16.87)
                                                ∂x
                                                     j¼1      j¼1
                                      T
           and s ¼ (s 1 , …, s n , s n+1 , …, s n+m ) . With these notations, neglecting the ECM remodeling, it is possible to summarize the
           governing equations as:
                                                          ∂u  ∂f
                                                                + s                                        (16.88)
                                                          ∂t  ¼  ∂x
                           ∂u
           where f ¼ f x,t,u,  and s ¼ s(x, t, u). Eq. (16.88) has sense if and only if we define suitable boundary conditions and
                           ∂x
           initial conditions. Boundary conditions are for each variable u i , i ¼ 1, …, n given by Eq. (16.2)or(16.3) and for each
           variable u i , i ¼ n +1, …, m by Eq. (16.5) or Eq. (16.6) except for the fact that we replace directional derivatives  ∂  by
                                                                                                             ∂n
                            ∂
           partial derivatives .
                            ∂x
              Finally, the initial conditions are
                                                               0                                           (16.89)
                                                          u ¼ u ðxÞ
                                                0
                   0
                       0
           where u ¼ C for i ¼ 1, …, n and u 0  ¼ S for i ¼ 1, …, m.
                   i   i                   n + i  i
           16.5.2.2 Weak Form
              The differential equation (16.88) with boundary conditions (16.2), (16.3), (16.6), (16.6) and initial condition (16.89)is
           a nonlinear parabolic differential equation in time and one space variable. We solve numerically this equation using a
           method based on a simple piecewise nonlinear Galerkin second-order accurate in space [47], which is compatible with
                                                                                                   1
           this kind of nonlinear equation and boundary condition. Multiplying the PDE by a test function ϕ 2H ð½α;βŠÞ and inte-
                                                                                                   0
           grating by parts in [α;β], we arrive to:
                                                              Z  β ∂ϕ    Z  β
                                                                            ϕQ dx                          (16.90)
                                           ϕðβÞfðβÞ ϕðαÞfðαÞ       f dx ¼
                                                               α  ∂x      α
                             ∂u    ∂u
           where Q ¼ Q x,t,u,       sðx,t,uÞ. As a test function we select
                             ∂t  ¼  ∂t
                                                                β  x
                                                          α
                                                                β  α
                                                        ϕ ðxÞ¼
                                                                x α
                                                          β
                                                        ϕ ðxÞ¼  β  α                                       (16.91)
           When using ϕ α , Eq. (16.90) becomes

                                                  Z  β  ∂ϕ         Z  β
                                                     f  α dx ¼ fðαÞ +  Qϕ  dx                              (16.92)
                                                                         α
                                                   α  ∂x            α
           and when using ϕ β , we get
                                                Z  β  ∂ϕ          Z  β
                                                   f  β dx ¼ fðβÞ +   Qϕ β  dx                             (16.93)
                                                 α   ∂x             α




                                          II. MECHANOBIOLOGY AND TISSUE REGENERATION
   325   326   327   328   329   330   331   332   333   334   335