Slurry infiltration and filter cake formation are critical for excavation surface stability in slurry shield tunnelling. Laboratory column tests are frequently adopted to study macroscopic infiltration. However, these tests, in a vertical orientation and confined by an impermeable cylindrical boundary, may not be a good representation of horizontal infiltration into an unbounded stratum in real-world tunnelling. In this paper, a more realistic slurry pressure balance (SPB) tunnelling model was created to overcome the limitations of simulated laboratory column tests. Coupled computational fluid dynamics (CFD)–discrete element method (DEM) numerical simulations were carried out to study the slurry infiltration into sand for this model and compare the results with the conventional laboratory column test model. Both models produced the same types of filter cakes. The SPB tunnelling model yielded larger infiltration distances and ranges, but a lower normalised permeability in the sand region in front of the tunnel face. The fluid pressure within this region dissipated much faster in the SPB tunnelling model, resulting in rapid velocity decreases and a faster infiltration process. Although the SPB tunnelling model is more representative of real-world tunnelling, the conventional laboratory column test is conservative, producing similar types of filter cakes over a longer timeframe.