Durability　of　concrete　in　engineering　structures　is　becoming　more　and　more　of　a　major　problem.　Research　into　such　problems　is　complicated　and　expensive,　however.　Developments　in　computer　technology　make　it　possible　nowadays　realistically　simulating　cementitious　materials　and　studying　its　pore　network　structure.　In　this　paper　use　is　made　of　the　dynamic　DEM　(Discrete　Element　Method)　denoted　HADES　(Habanera＇s　Discrete　Element　Simulator),　introduced　earlier,　and　a　new　vector-based　hydration　simulation　system,　XIPKM　(Extended　Integrated　Particle　Kinetics　Model)　.　Inspired　by　developments　in　robotics,　a　pore　exploration　system　is　developed　in　DraMuTS　(Double　Random　Multiple-Tree　Structuring).　The　tree　structure　inside　pores　governs　pore　topology,　including　pore　continuity.　For　determination　of　pore　geometry　such　as　pore　size,　a　second　uniform　random　point　system　is　used.　In　each　of　these　points,　star　volume　measurements　govern　local　pore　size,　while　the　collected　data　allow　constructing　a　volume-based　pore　size　distribution.　The　method　is　illustrated　by　means　of　data　pertaining　to　Portland　cement　(PC)　paste　and　rice　husk　ash-blended　PC.　This　approach　is　demonstrated　providing　all　necessary　pore　network　information　relevant　for　durability　estimation.