Cascade Saturates When Cooling Water Warms Up
A product cooler uses an outlet temperature controller cascaded to a cooling-water flow controller. Design basis: 30,000 lb/h of process fluid, cp = 0.55 Btu/lb·°F, cooled from 250 °F to 180 °F. Cooling water enters at 85 °F, is allowed to return at 105 °F, and design flow is 115 gpm through a valve that passes a maximum of 130 gpm at 100% output. The cascade ran well all winter. In July the cooling-water supply temperature rises to 95 °F, and a downstream constraint now limits the return to 110 °F. The temperature controller output has been pegged at 100% for six hours, the outlet temperature has drifted to 195 °F, and the flow loop is at its maximum. The console operator keeps increasing the primary controller gain; nothing improves.
1. Show by calculation that this is a capacity limit, not a tuning problem.
Read the worked answer
Duty required
Q = ṁ cp ΔT = 30,000 × 0.55 × (250 − 180) = 1,155,000 Btu/h
Design check (winter): 115 gpm ≈ 57,500 lb/h at 500 lb/h per gpm → 57,500 × 1.0 × 20 °F = 1,150,000 Btu/h. Design is consistent.
July requirement
Allowed water ΔT is now 110 − 95 = 15 °F.
Required flow = Q / (cp ΔT) = 1,155,000 / (1.0 × 15) = 77,000 lb/h ≈ 154 gpm
The valve can pass only 130 gpm - a ~16% shortfall. The loop is saturated; the primary output at 100% is correct behaviour, and no gain value can create flow that the valve cannot pass. Increasing gain on a saturated controller only worsens windup.
Cross-check against the observed temperature
At the maximum 130 gpm (65,000 lb/h) and the 15 °F limit: Qmax = 65,000 × 15 = 975,000 Btu/h. Outlet = 250 − 975,000/(30,000 × 0.55) = 250 − 59.1 = ~191 °F
Observed 195 °F is a few degrees worse, consistent with normal fouling/approach loss on top of the capacity shortfall. The numbers line up.
2. What would you verify on the plant before concluding, and what would you tell the operator to do with the controller right now?
Read the worked answer
Verify before concluding
- Is the valve actually at 100%? Compare controller output to actual valve position on the positioner/limit switch - a stuck or mis-calibrated positioner, low air supply, or a hand-switch left in the wrong mode mimics saturation.
- Is the flow reading real? Check the flow transmitter against the pump curve / a second indication.
- Fouling. Compare the current approach (process outlet minus water outlet) with the clean design approach. Design: 180 °F process out vs. 105 °F water out ⇒ 75 °F approach. If the current approach has grown well beyond design, cleaning the tubes buys back capacity and may close part of the gap.
- Return-temperature constraint. Confirm the 110 °F limit is real and not a conservative console setting.
Right now
- Put the primary controller in manual/cascade-open or otherwise stop the integral from winding up - at 100% output with a persistent error, the integral term is accumulating a large negative-limit backlog that will cause a violent overshoot when conditions change.
- Verify external-reset-feedback / anti-windup is configured on the primary.
- Raise a high-temperature alarm and inform production that the cooler is over its capability.
3. List the realistic options to restore the temperature, and state which you would pursue first with justification.
Read the worked answer
Options, roughly in order of cost/speed
- Reduce throughput or inlet temperature. Dropping process flow to Q/(cp ΔT) capacity: at 975,000 Btu/h available, maximum flow = 975,000/(0.55 × 70) = 25,300 lb/h - about 16% rate cut. Immediate, but costly.
- Clean the exchanger / check for fouling and air binding. Cheapest capital-free gain and usually the first thing to check; recovering even half the lost approach may be enough.
- Reduce the return-temperature constraint margin if the downstream limit permits - each extra degree of allowed rise cuts required flow by roughly 6–7%.
- Add cooling capacity: a trim cooler, a larger valve and water line (130 → 155+ gpm), or chilled water to restore ΔT.
- Lower the water supply temperature by fixing the cooling tower, if the 95 °F is itself abnormal for the system.
First choice
Clean/inspect the exchanger and verify the valve really is fully open - the required shortfall is only ~16%, which is within reach of restoring design approach. In parallel, take the rate cut so the temperature is back inside limits while the mechanical work is done. Adding cooling capacity is the right long-term answer if July conditions are the new normal and the constraint is genuine; that decision should be made from a proper rating calculation, not from the console.