How to set the initial turbulence of the turbulent combustion in the fan-sirred bomb？
 

Hello everyone!
I recently using LES to simulate the turbulent combustion in the fan-stirred chamber. But how to set the initial turbulence field confused me a long time. So i post my question and hope you can give me an advance. My question is as following：
Most of the simulation of turbulent combustion within the fan-stirred chamber
was studied based the MK2 fan-stirred chamber of University of Leeds， which inner diameter is 380mm， and turbulence was generated by four fans in the chamber（Fig1 and Fig2 ）.

Because of the high computational domain, most researchers simplfy the domain as a cube, whose edge is 100mm or 160mm，and set all the boundary as the periodic boundary.I recently also use the method to simulate the turbulent combustion in the fan-stirred chamber. Presently I
adopt two methods to calculate the domain setting of initial flow field.
1.Using 3 pairs of translational periodic boundary, set pressure gradient.
after calculation, the calculation of velocity distribution of boundary is different the profile of the flow simulation of the fan-stirred chamber so think of the current calculation results may have a problem.
2.The second method is to simulate the flow field of the fan-stirred chamber
, cuting the six face as the profile of the domain（100mm3 cube），But i found that after the calculation, the mass flow of the center face of the fan-stirred chamber is 0， and the simplified model is not 0. I dont know where isthe problem.
In addition, the ming-bo sun professor used the isotropy turbulent quickly generate method to simulate the turbulence, but the details is unknown. Scholars have also used the Uhlenbeck - Ornstein, a statistical method, but due to the related literature online do not have the resources nor too understand.
Looking forward to your help，thank you！