hi users, I am working in CFX 5.6. I have a project in which a water tank is translated by means of a truck. How can I model the movement of water when the truck is accelerating or deaccelerating? i took a part of the tank in which a baffle whith two opennings are existing. I tried to use transient simulation type. In the boundary condition, I gave the inlet an exponential expression as an input for the inlet speed. Is there any rong with this simulation? Thanx

It depends upon what you are trying to get from the solution. If you are looking at say forces on the walls of the tank then you would have to take into account the effects of added mass, and to do this you would need to simulate the actual movement of the tank. The movement of the tank can be done using the mesh movement options accessed through user fortran. For this you would need to contact the CFX rep. Conversely you could wait until 5.7 comes out which may be at the start of April. In 5.7 there will be a GUI for the mesh movement I believe (not 100% sure). Good luck. Bob

What do you mean by mesh movement and from where can I find this option?? Thanx in advance

in reality the tank moves, so moving the water. To model this in CFX you would need to move your mesh with the same accelerations and decelerations as the tank would experience. There may be assumptions you could make to try and generate a less complex method to solve your problem but it probably would not be as accurate. The best place to find out about this would be to talk with your CFX rep, or call the support guys. Bob

I think if the domain has an acceleration a (vector), it is the same as being under a force field minus a. So it should be enough to add the vector minus a, as a function of time (I suppose it is the case), to the gravity vector. If the domain rotates, the gravity vector and the minus a vector move relatively to it, and their x,y,z components have to account for this. If there is the need to take in account the rotation acceleration, another force field, with modulus and direction function of the position relatively to the center of rotation, and possibly function of the time, can be added. At least this sounds logical to me. Rui