How to chose the cycle of convergence

Last question one may thought before running a simulation would be: “How many cycles should I let the simulation run?”. Well, the simple answer is: “It depends”. In fact, it depends on a lot of things: engine type (turbocharged engine usuially need more time for getting a steady solution (convergency)), engine geometry (the ratio between engine displacement vs. intake and exhaust manifold sizes), flow initial conditions (if they are close to the final solution, the number of cycles needed for a converged solution is lower).
Here you’ll find some hints on how to choose the number of cycles a simulation should run in order to get a good solution (I’m assuming it is an engine steady running condition what you want to run, not an engine transient behavior):

• First time I run a simulation I use 100 cycles. OK, the simulation will take a lot of time but it is only one simulation.[/*:1vssc5ki]
• I always check the averaged air mass flow in the first intake duct and in the last exhaust duct (the comparison is simple: intake-mass + fuel = exhaust-mass). These values can be seen in the MED.DAT file but you have to include them before running the simulation. If the previous equation does not fit, do not blame the code, it means that intake or exhaust manifold flow conditions are not steady yet.[/*:1vssc5ki]
• Another way to see if the solution has converged is to check if averaged pressure values in different engine locations have achieved a steady evolution.[/*:1vssc5ki]
• If the engine is turbocharged, it is good to check and compare the average turbine and compressor work (in the MED file; again be sure to include them before running the simulation). Both works should be very similar (in pure theory they should be equal). If turbine work is greater/lower than compressor work you should detect that the turbo speed is increasing/decreasing. So turbo speed is also a good variable to check if a simulation has converged.[/*:1vssc5ki]
• Finally, you have to decide in which cycle the previous comments are well satisfied and choose it as a good value for the number of cycles. For example, I’ve attatched a plot containing intake and exhaust mass flows and it is observed that 20 cycles is enough for arriving to converge.[/*:1vssc5ki]
• Remember that a number of cycles which is appropriate for a given engine running condition may not be good for another one (mainly when changing engine speed or turbocharger operation).[/*:1vssc5ki]