Density. CO2 is carried as a dense fluid. Its density comes from the Peng-Robinson equation of state, P = RT/(v-b) - aα(T)/(v(v+b)+b(v-b)), with a Peneloux volume shift, from the NIST critical data (Tc 304.13 K, Pc 7.377 MPa, ω 0.2239). A cubic equation of state is within about 6 % of the NIST tables (check below), good enough for hydraulics.
Compression. Gas from the capture plant (0.11 MPa) is compressed in 4 intercooled stages to the critical pressure, W = n (γ/(γ-1)) (RT/M) [(p2/p1)(γ-1)/(nγ) - 1] / η, then a liquid pump lifts it to the line pressure, W = ΔP/(ρη). The total of about 95 kWh per tonne is in the published range and replaces the old unexplained "kWh/yr/t" slider.
Pipeline. Pressure drop along the line: dP/dx = f ρ v²/(2D) with the Swamee-Jain friction factor and the local density; whenever pressure falls below 8.6 MPa (so the CO2 stays dense) a booster pump station restores the inlet pressure. A velocity limit of 3 m/s caps the flow of a narrow pipe.
Storage. Injection equals the smallest of capture, pipe capacity and well capacity. Stored mass grows as Q t. Reservoir pressure is P0 (hydrostatic at 1.5 km) plus an over-pressure that grows in proportion to the filled fraction of a closed, compressible aquifer, up to half the margin to the fracture pressure (15.8 kPa/m). Net abatement subtracts the CO2 of the electricity used (0.4 t/MWh).
Simplifications: cost is levelised operating cost only (no capex); capture energy is inside the $/t figure; constant dense-phase viscosity 80 µPa s; isothermal 15 C line at ground temperature; the reservoir is a lumped pressure model without plume buoyancy or leakage, so no leak probability is claimed (the old "risk index" that rose with monitoring spend was wrong and has been removed). The pipe is drawn above ground for visibility. The plume lens is a schematic, not a flow simulation.
Running numerical checks...