Stuck Cooling Valve on an Exothermic Batch Polymerization
A 5,000-gal jacketed batch reactor is running a 6,000 kg emulsion polymer batch with an adiabatic temperature rise of 180 K. Reaction onset, where self-heating outruns the jacket, is 120 C. At 09:14 the reactor is at 95 C and the jacket cooling-water control valve fails, sticking at 20% open. Over the next 8 minutes the reactor climbs from 95 C to 108 C, and the heating rate is still increasing at roughly 4 K/min. Vessel design pressure is 150 psig. The relief valve was sized for an external fire, and its outlet feeds a header shared with two other reactors. You have 400 kg of 5 wt% aqueous inhibitor on hand, injectable through a dip tube, and 500 kg of 10 C quench water available. The batch heat capacity is about 3.0 kJ/kg-K.
1. Estimate the time available before onset and explain why the margin is smaller than the arithmetic suggests.
Read the worked answer
Onset is at 120 C and the reactor is at 108 C, so the gap is 12 K. At 4 K/min that is 3 minutes, and that is the optimistic number.
The reaction is autocatalytic in temperature: the rate is still accelerating, so the 4 K/min is a lower bound on the next interval, not a constant. Self-heating roughly doubles for every 10 K, so the last few degrees before onset will pass far faster than the first few. Realistically you have 2 to 2.5 minutes of useful response time.
Treat the 3-minute figure as a deadline for the first action, not for the whole response. After onset the jacket can no longer remove the heat, the temperature rises toward the maximum adiabatic value of about 95 + 180 = 275 C, and the vessel pressure follows the vapor pressure of the monomer and water at that temperature. This is why the action has to be something that stops the chemistry, not something that adjusts the cooling, because the cooling is already gone.
2. Compare the inhibitor injection and the water quench quantitatively, and state which you would use first.
Read the worked answer
The quench is far too small to solve the problem. The batch heat capacity is 6,000 x 3.0 = 18,000 kJ/K. Warming 500 kg of water from 10 C to 108 C absorbs 500 x 4.18 x 98 = 204,800 kJ, which cools the batch by 204,800 / 18,000 = 11.4 K. That buys under 3 minutes at 4 K/min, and it dilutes the batch, adding volume and possibly destabilizing the emulsion. It is a delay, not a stop.
The inhibitor removes the heat source rather than the heat. At 400 kg of 5 wt% solution you have 20 kg of active material, which is meaningful against a 6,000 kg batch if the injection point actually reaches the bulk. The catch is mixing: a dip tube in a viscous emulsion may channel, so inject while the agitator is running and verify the agitator is still on.
Order of operations: block in the cooling water manually at the valve bypass if one exists, inject inhibitor, and start transferring to a dump tank if the level permits. Keep the quench in reserve for cooling the vent line or the dump tank, not the reactor.
3. The plant manager argues that the relief valve exists for exactly this situation. Evaluate that position.
Read the worked answer
It is wrong on three counts.
First, the sizing basis. The valve was sized for an external fire, which assumes the relieving fluid is vapor with a known latent heat and a fire heat input computed from wetted area. A runaway polymerization produces gas dissolved in a foaming, viscous liquid. The relieving stream is likely two-phase with a much higher mass flux and a much higher effective heat release than the fire case, so the certified capacity does not apply.
Second, the disposal system. The outlet goes into a header shared with two other reactors. A large two-phase release can back-pressure that header above the other vessels' set pressures and can carry monomer or polymer into the flare, where a plugged or burning header becomes the incident.
Third, the hierarchy of controls. The relief valve is the last line of defense, and its design basis has already been invalidated by the scenario. Injection of inhibitor and inventory transfer are the controls; relief is what protects the vessel if those fail. Relying on it converts a recoverable upset into a release.