This simulator models a vessel that fills and drains while two independent level technologies watch it: a bottom-tap hydrostatic DP transmitter with a selectable dry or wet reference leg, and a roof-mounted radar sensor that times an echo off the liquid surface. Adjust inflow, outflow, vessel geometry, gas pressure, liquid density, reference-leg configuration and sensor faults, then compare what each technology reports against the true liquid level.
• A real-time 3D cutaway of the vessel shell, liquid volume and surface, inlet nozzle, bottom outlet/drain valve, bottom DP tap and capsule, dry/wet reference leg, and roof radar antenna and echo path, with home view, focus-selected-part, full-enclosure/cutaway toggle, exploded view, auto-rotate, expand and show/hide numbered labels matching a companion diagram. • Twelve live controls: initial level (applies at simulated time zero), tank cross-sectional area, inlet flow, outlet flow, actual liquid density, calibrated liquid density, closed-vessel gas pressure, vessel reference mode (vented / closed-dry-leg / closed-wet-leg), wet reference-leg density, a compensate-wet-leg-head checkbox, displayed transmitter (hydrostatic DP or radar time-of-flight), and radar condition (healthy echo / echo lost / false echo at 0.6 m). • Play/pause, single-step (0.1 s) and larger-step (1 s) time controls, plus a playback-speed selector (10x slow motion, real time, 10x faster, 1 minute per second). • Start/stop trial, fill (set inlet to 20 L/s) and hold (stop inlet and outlet) actions, with a live "what is happening" sequence narrative, operating state, switch-state tokens and a scrollable cell-readings table. • Eleven live metrics: actual liquid level, contained liquid volume, bottom tap pressure, reference port pressure, raw differential pressure, selected sensor estimate, selected transmitter current, valid reading error, radar round-trip travel time, measurement validity flag, and cumulative overflow. • A Curves & measurements tab with two charts (actual level vs. selected estimate, and bottom/reference/differential pressures), the full model equations, and snapshot measurements. • An Experiments tab with four guided scenarios (fill the vessel, density mismatch, uncompensated wet leg, lost radar echo), a model-verification bench of independent automated checks, and a timestamped event log with a copyable trial report. • A Learn & assess tab with guided lessons, a knowledge-check quiz and a written scope/reference statement.
The DP transmitter infers level indirectly, from Pbottom = Pgas + ρgh: it measures a pressure and converts it back to a height using an assumed liquid density. If the calibrated density doesn't match the actual liquid density, the DP-inferred level is biased even though the transmitter itself is working correctly — the simulator's density-mismatch experiment shows a real 2 m column reported as 1.6 m for exactly this reason. Radar instead times a round-trip echo to the surface and back (d = c·t/2), so its distance estimate does not depend on liquid density at all, making it a useful cross-check against a DP reading that might be affected by a density change.
On a closed vessel, gas pressure sits on top of the liquid and would corrupt a bottom-only pressure reading, so a reference leg is used to cancel it: ΔP = ρgh, because the same headspace pressure appears at both the bottom tap and the reference port and subtracts out. A wet reference leg complicates this further by adding its own hydrostatic head at the low port (ΔPwet = ρgh − ρwet·g·Hreference) — if that known offset isn't compensated back in, the transmitter can even show a negative differential pressure while the vessel actually holds liquid.
The metrics separate a sensor's raw output from whether that output should be trusted: the model reports "valid" and "valid reading error" as 0 whenever a fault makes the true measurement unavailable, rather than silently returning a plausible-looking but wrong number. A lost radar echo, for instance, drives the transmitter current to a representative fail-high value while the displayed estimate is explicitly flagged invalid — the point being that an invalid, fail-high signal must not be mistaken for a genuinely empty tank reading zero.
The full inventory and sensing equations are shown in the equations panel: dh/dt = (Qin − Qout)/Atank for the level dynamics, the bottom/reference/DP relations above, and d = c·techo/2 with hradar = 5.5 m − d for the radar geometry. The stated scope notes a uniform 5 m level span with a 5.5 m sensor elevation and a fixed 5 m wet-leg height, vented mode fixing gas pressure to atmosphere, and no foam, boiling, dielectric-strength or full radar signal-processing model — radar faults are explicit educational scenarios, not a physical antenna simulation.
The DP transmitter infers level from measured pressure using Pbottom = Pgas + ρgh, then converts that pressure back to a height using the calibrated density. If actual liquid density differs from the calibrated value, the inferred level is biased even with a perfectly working transmitter. Radar instead times a round-trip echo to the liquid surface, so its distance estimate is independent of liquid density altogether — the simulator's density-mismatch experiment shows this contrast directly.
A closed vessel's headspace gas pressure acts on both the bottom sensing tap and the reference port equally, so it cancels when the transmitter computes the differential pressure between the two: ΔP = ρgh. This is why closed-vessel level measurement uses a reference leg rather than a single absolute-pressure tap.
A wet reference leg is filled with liquid and adds its own constant hydrostatic head at the low port, per ΔPwet = ρgh − ρwet·g·Hreference. If that known offset is not added back in through compensation, the transmitter can report a negative differential pressure — and therefore an implausible negative level — even while the vessel genuinely holds several meters of liquid.
A lost radar echo drives the transmitter to a representative fail-high current, and the displayed level estimate is explicitly marked invalid rather than shown as zero, so the fault is distinguishable from a genuinely empty vessel. The model uses a uniform liquid density, constant cross-section, ideal DP taps and ideal radar propagation with a fixed 5.5 m sensor elevation and 5 m wet-leg height, and does not simulate foam, boiling, dielectric strength or real radar signal processing.