EIGENROOMLAB WORKSHEET 36ANSWER KEY · FOR INSTRUCTORSWater tank
Replace water lost to a drain
Goal
Compare a fixed pump command with PI level control when an extra drain opens.
Setup
Lab eigenroom.com/lab/tank?lesson=tank-disturbance · short link eigenroom.com/tank-disturbance
Open it on a phone or a computer; there is nothing to log in to or install. The link loads run A's settings:
- Tank area A
- 0.1 m²
- Outlet coefficient c
- 0.003 m⁵ᐟ²/s
- Linearization height h₀
- 0.5 m
- Initial height
- 0.5 m
- Target height
- 0.5 m
- Pump control
- PI + q₀ bias
- Flow step Δq
- 0 m³/s
- Pump limit
- 0.005 m³/s
- Kp
- 0.004 m²/s
- Ki
- 0.00015 m²/s²
- Anti-windup
- On
- Extra drain from 30 s
- 0.0008 m³/s
- APreset, exactly as the link opens it.
- BChange only: Pump control → Manual step about q₀.
- CChange only: Ki → 0.0003 m²/s². (ext.)
1. Predict
An extra drain opens at 30 s. Will fixed pump flow and PI react in the same way?
PI measures the level error and adjusts pump flow; a fixed command cannot replace the changed loss automatically. The q₀ bias already balances the original operating point. I supplies the additional sustained correction. A pump limit can make a target unmaintainable, and no controller can refill faster than the available inflow.
2. Observe
| Quantity | Where to read it | A | B | C ext. |
|---|---|---|---|---|
| Nonlinear height at t = 60 s (m) | Readout at 60 s | 0.41293 m | 0.32621 m | 0.43176 m |
| Nonlinear height at t = 120 s (m) | Readout at 120 s | 0.48142 m | 0.21357 m | 0.50881 m |
| Pump flow at t = 120 s (m³/s) | Readout at 120 s | 0.0029632 m³/s | 0.0021213 m³/s | 0.002943 m³/s |
| Integral contribution at t = 120 s (m³/s) | Readout at 120 s | 0.00076757 m³/s | 0 m³/s | 0.00085695 m³/s |
| Final target error (m) | Results below the plots | 0.018576 m | 0.28643 m | -0.0088108 m |
Values computed by State-feedback, tank and RLC engine 1.0.0 at the settings above. A replay readout can differ in the last digit if the slider is not exactly at the stated time; values read from a plot are approximate.
3. Explain
Compare the pump flow at 120 s. Which run replaced the lost water?
PI raised the pump flow to 0.0029632 m³/s, close to the 0.0029213 m³/s needed after the drain opens; the manual pump stays at 0.0021213 m³/s, so its level keeps falling, to 0.21357 m.
Run A is still below its target at 120 s. Is that a steady-state error?
No: the integral term is still growing (0.0002875 m³/s at 60 s, 0.00076757 m³/s at 120 s). With this small Ki the recovery takes longer than the recording; integral action removes a constant error only over time, at a rate set by Ki.
Extension (optional)
Raise Ki to 0.0003 m²/s² (run C). How close to the target does the level end?
Closer, and slightly above it: the final error is -0.0088108 m (negative means above the target), against 0.018576 m for run A, and the level at 60 s is 0.43176 m instead of 0.41293 m. A larger Ki rebuilds the needed flow faster and can carry the level a little past the target.