Installing a mechanical seal on a reactor takes four steps, in this order: preparing the vessel and checking the seal and shaft, fitting and setting the seal on the nozzle flange, connecting the barrier and cooling lines, and proving the seal under pressure before the agitator runs.

Most process reactors run a double mechanical seal with pressurised barrier fluid, and the steps below assume one. A single seal is used only on benign, non-toxic duty, where the barrier step becomes a quench or cooling-water connection.

Step 1: Prepare the Reactor, the Seal and the Shaft

Three things are settled before the new seal leaves its box: the reactor is safe to open, the seal is the right one for this vessel, and the shaft and nozzle flange are within the seal’s limits.

Isolate the Reactor and Take Out the Old Seal

Lock out the agitator motor and the valves to and from the vessel, relieve the pressure, then drain, clean and cool the reactor.

Secure the agitator shaft as the agitator maker instructs before the coupling is released, so the shaft and impeller cannot drop into the vessel. Then remove the coupling or drive to reach the seal. A spacer coupling lets the seal come up the shaft without moving the gearbox; otherwise the drive is lifted off. Some agitator makers want the whole agitator lifted off its tank flange before anything is dismantled.

Take the old seal off the nozzle flange and up the shaft, holding the rotating and stationary parts so neither falls when it comes free.

Check the New Seal Against the Reactor Duty

Check the new seal against its assembly drawing and against the vessel. The shaft diameter, the vessel flange bolt size and bolt circle, and the distance from the flange to the coupling or drive must all match the drawing, and so must the materials and the piping connections.

Then check the duty. A reactor seal is ordered for a stated barrier fluid, maximum vessel pressure, shaft speed and temperature, and it is not used outside them.

Out of the box, look for unmarked faces, O-rings seated and undamaged, and the setting clips still in place on a cartridge seal.

Check the Agitator Shaft and the Nozzle Flange

Measure the agitator shaft at the seal position and judge it against the mixer seal’s limits, never against pump limits. An article by Flowserve’s mixer-seal engineers puts numbers on the gap: API 682 limits a pump shaft to 0.002 in TIR at the seal faces, while 0.03 to 0.150 in TIR is the normal range on a top-entry mixer.

Shaft radial runout at the seal (in TIR)
Pump shaft, API 682 limit at the seal faces
0.002 in
Top-entry mixer shaft, normal range
0.03–0.15 in

Each agitator maker sets its own tolerances, so the limits that apply are the ones on your seal’s drawing. For example, one OEM’s top-entry cartridge seal requires:

CheckLimit on one OEM’s mixer cartridge seal
Vessel flange face square to the shaft axis0.25 mm (0.010 in) FIM
Vessel flange concentricity to the shaft, static0.50 mm (0.020 in) FIM
Vertical shaft movement (axial end play)0.61 mm (0.024 in) FIM
Shaft diameter±0.025 mm (±0.001 in)
Shaft surface finish0.8 µm Ra or better
Vessel flange face finish1.6 µm Ra or better

The bearings, drive and coupling must also be in good condition. The dial-indicator set-up is the same as measuring runout on a pump shaft; only the limits change.

Step 2: Install the Seal on the Reactor

On a reactor the seal is set with the shaft in its running position. A cartridge seal is lowered onto the nozzle flange but not bolted, the drive is refitted and the shaft brought to its final operating position, and only then are the flange bolts tightened, the set screws locked and the setting devices removed.

A pump cartridge seal is bolted to the seal chamber first. On an agitator, refitting the drive and releasing the shaft restraints move the shaft, and a seal locked before that is set to the wrong position.

Clean and Lubricate

Clean the shaft, the vessel flange face, the gland and the seal bore, and break every sharp corner on shaft steps, shoulders, keyways and threads the O-rings must pass over.

Lubricate the O-rings, and the shaft along the length the sleeve O-rings travel, with a lubricant compatible with both the elastomer and the process. Silicone grease usually suits, but verify it against the process first.

The faces stay clean and dry. Wipe them with a lint-free cloth and a quick-drying solvent or alcohol, and do not lubricate them unless the seal drawing says so.

Fit the Seal to the Nozzle Flange and Shaft

Lower a cartridge seal onto the nozzle flange as a unit on its lifting eyes, setting-device end toward the drive, and leave it unbolted. Take care as the sleeve O-rings pass over keyways or threads.

Refit the coupling or drive, release any shaft clamps or restraints, and complete the adjustments that put the shaft in its final operating position. Then run the flange studs down evenly, crosswise in stages, to the torque on the seal drawing.

Tighten the sleeve set screws uniformly to the drawing torque last.

On a component seal, the stationary ring and gland go to the flange first, with the ring level and square to the shaft. The rotating unit slides down the shaft over a guide sleeve, is never struck into place, and is locked after the shaft is in its running position.

Set the Seal Working Length

On a cartridge seal the setting clips or plates hold the working length, and they are disengaged only after the flange is bolted and the set screws are tight. Secure them clear of the sleeve so they cannot fall back onto rotating parts.

On a component seal, the setting distance comes from the assembly drawing and is measured with the shaft where it will be in operation, including its thermal growth relative to the vessel. The measurement itself is the same as setting compression on a pump seal.

Check by Hand Before Connecting Anything

Before any line is connected, turn the shaft slowly by hand. It should turn freely with no rubbing, and the faces must not be stuck together.

Then confirm that nothing still locks the seal: setting clips or plates disengaged and secured, centring tabs removed, shaft clamps and transport locks off, and any distance pieces fitted as the drawing shows.

Step 3: Connect the Barrier Fluid and Cooling Lines

Before the vessel is pressurised or the shaft turns, a double seal’s barrier system must be connected, filled, vented and at a pressure above the process pressure at the seal. On a single seal, the quench or cooling water is connected and on instead.

Pipe the seal pot to the barrier inlet and outlet ports and the cooling water to the cooling ports marked on the seal drawing, and route the lines so they leave no air pockets.

The margin comes from the seal maker’s data sheet, which states a minimum operating barrier pressure. The seal’s barrier pressure rating is the upper limit.

The sequence is the same as filling and pressurising a pump seal pot. Connect the nitrogen supply and check it for leaks, fill the barrier fluid, vent the lines and the seal, then raise the nitrogen to the required pressure and leave the supply on.

Turn on the cooling water to the pot coil. Seal makers’ guidelines keep the barrier fluid below 80 °C, measured close to the pot. A cooling flange on the seal itself is a separate item, used when the product runs above about 200 °C to keep the O-rings within their temperature limit.

Leak-test the pot and its piping before relying on them. One seal maker tests the complete system at 1.1 times its maximum working pressure for 5 minutes.

Step 4: Pressure Test and Start Up the Reactor

Prove the seal twice before anything is fed to the reactor: first static, under pressure with the shaft still, then turning at low speed.

Static Pressure Test

Pressurise the vessel, usually with nitrogen, with the barrier system already at its pressure. Hold it while you check the gland, the flange joint and the seal ports, and watch the pot level and pressure. A falling pot level or a leak at the gland means the seal comes out before the agitator runs.

The test pressure with the seal fitted is capped by the seal’s static rating, not by the vessel’s code test pressure. A code hydrotest above the seal’s rating is done before the seal is fitted.

First Start-Up and Run-In

Check the rotation direction against the nameplate and confirm that barrier pressure and cooling water are on. Start the agitator at low speed, with a soft starter or frequency converter where one is fitted, then raise the process pressure slowly.

In the first minutes, check that the barrier is circulating: the return line to the pot should run warmer than the supply. If no flow shows after several minutes, stop, look for blockages or trapped air, and vent the lines.

For the first 2 to 3 hours, until the system reaches its equilibrium temperature, watch barrier pressure, temperature and pot level, seal-area temperature, vibration, noise and leakage at the seal and piping. There is no dry run-in: the faces never turn without barrier or process fluid between them.

Stop and find the cause of the leak if the barrier pressure becomes unstable, the pot level moves quickly, the seal area keeps heating, or leakage appears at the seal.