I have modelled an unbaffled stirred tank reactor using the MRF technique and the realizable k-e turbulence model.

I have then used the discrete phase model to add an injection of a single inert particle. The Manual section 19.1.3 states that particle tracks in this scenario are not meaningful, and recommends that the text command "define/models/dpm/tracking/track-in- absolute-frame" is used.

I enabled this command and then plotted the residence time against local shear rate (which I defined as a custom field function).

However, I have written six separate files of residence time and local shear rate. Each time I have different results. The residence time varies on each occasion (11.99s, 12.38s, 16.09s, 16.44s, 7.19s, 8.48s)

In the in the Define ->Models->Discrete phase Panel I have set the Max No of steps to 15000 and the length scale to 0.01m. I have not checked the Interaction with Continuous Phase Box or the Saffman Lift Force box. The Drag law is set to spherical.

The Solver Panel is set as Segregated, Implicit, 3D, Steady, Absolute velocity formulation.

The injection is just one particle (diameter 1e-6m, density 1000 kg/m3). The fluid is water (density 1000 kg/m3).

The values of the local shear rate (as defined using a cutom field function) are generally consistent between the different sets of results (i.e. they vary across the same size range), but the values occur at different particle residence times and the residence times are different for each set of results. Why am I getting these inconsistent results?

There is an additional term which represents the turbulence effect on the particle moving. If I understand your problem correctly, your inconsistent result should be caused by this term where the randomness is added. For a single particle, its residual time could change significantly especially when there are some eddies in the solved mean flow field. I suggest you calculate the averaged residual time by adding more particles. Hope it helps.

Jun