Take the Pressure Off Loading

Readers suggest how to avoid risks during upsets.

We use a centrifugal pump for tank truck loading. It's driven by a standard fixed-speed induction motor operating at 3,600 rpm. We have no controls on pump flow. The pump runs out on the pump curve during tank loading. An upset condition could cause a very high pressure at the pump suction. The piping system can handle the pressure but we have a concern about relief valve capacity downstream of the pump. The relief valve and downstream piping were designed based on flow being limited by pump capacity. But suction pressure can rise so high that the flow rate, if the pump weren't present, would exceed pump run-out capacity. Can the pump motor be forced to speed up to accommodate a higher flow rate by high pump suction pressure? Or will the motor and pump act as a brake on the system?

If, during an upset, the suction pressure increases so high that the pump is not necessary, the fixed-speed induction motor will run essentially at the same speed (3,600 rpm). The pump impeller will add kinetic energy and the volute will convert the kinetic energy to head pressure. The discharge pressure will be the sum of the suction pressure and the head developed by the pump. In hydraulic sense the pump cavity, the block valve and the piping represent the net resistance to the flow. The excessive discharge pressure can cause severe splashing at the tank and may pose a safety risk.

Consider the following improvements to reduce this risk:

1. Install a relief valve on the suction side. There is one present at the discharge side. However, if the upset is instantaneous, then the relief valve may not be much help — you need to make sure piping (suction and discharge) is adequate for the expected pressure.

2. If feasible consider throttling the discharge valve all the time; of course, this will slow down loading of the tank truck. However, it will help you keep on the pump curve operating range (the pump won't run off the curve). During an upset, however, you will have higher pressure than what you get with the valve wide open.

G.C. Shah, HSE project manager
Mustang Engineers, Sunnyvale, Calif.

As the suction pressure rises it will force flow through the pump and the motor current will fall since the suction pressure is doing the work originally performed by the motor. If the suction pressure continues to rise the flow will increase and start free-wheeling the pump shaft above its nominal speed. At that point the pump motor will become a generator and will start producing current in an attempt to slow down the pump. So the short answer to your question is, "Yes, the motor will act as a brake until it finally blows the overloads in the starter and trips off line. At that point the pump will likely free-wheel and do little to restrict the flow into the downstream piping."

Hunter Vegas, senior project manager
Avid Solutions, Inc., Winston Salem, N.C.

A three-phase induction motor will only spin at (or very near) the rated rpm when energized. If shaft work required actually does go below zero, the motor will act as a brake and produce electrical power rather than consume it.

Chris Rentsch, process engineer
Dow Chemical, Midland, Mich.

As the pump inlet pressure increases due to upset, the liquid flows to discharge side at increasing rate. (It was assumed pump discharge side is at lower pressure compared to suction side due to safety valve opening.) This is the case of a hydraulic turbine, where the liquid head drop across the impeller converts into mechanical work. This is in addition to the electrical energy supplied to the pump motor as there were no controls specified. Therefore in such a case, the pump shaft can acquire very high speeds depending the differential pressure. Depending on the properties of fluid handled, high pressure fluid depressurization across the pump may cause severe cavitation and damage to the pump.

C.C.S. Reddy, lead process design engineer
Singapore Refining Co., Singapore

I suspect that the pump is being operated by a variable frequency drive. I assume the pump is a centrifugal radial type. Higher speeds are probably possible — but why would you want to go there? Pump wear is to a power factor of three. Going above 3,600 rpm increases wear on bearings and seals. Perhaps this pump can run at a higher speed but this approach should be discouraged. Instead, install an orifice plate or a series of plates for extreme pressure. A restrictive orifice (RO) will bring the pump back on its curve during unloading. Size the RO ΔP using the difference between the extreme pressure and the design pressure. Install the RO upstream of the relief valve to prevent over-pressurization of the piping.

Think about the possibility of other equipment damage. Inspect the pump case and all equipment upstream of pump discharge to assure they will survive the higher pressure. It helps to remember that pressure ratings are based on the probability of failure. A valve rated for 150 ANSI may survive 200 psi without a leak but the odds of leaking or even bursting increase exponentially above its rating. If the expected pressure exceeds ratings, why not install a relief valve upstream of the pump in the suction piping?

Will the pump act as a brake on the system? I doubt it. The shaft will turn counterclockwise. The impeller may eventually spin off. Check to see if the impeller is keyed in such a way to prevent this. Turning the shaft will not be detrimental to the motor, although there could be a risk to turning a motor into a generator.

Dirk Willard, senior process engineer
International Steel Services, Inc., New Caledonia

We had six railcars in a propylene loading bay (Figure 1). One of the tankers was over-filled and was vented to fixed piping via hose at the loading station; the fill valves on the car and the loading arm were open. The other cars' valves and their loading valves on the fixed piping were closed per standard procedure. Sometime during the night two additional cars were brought down the line to be filled, pushing the first six cars towards the exit and blocking the entrance track point. The locomotive crew -- believing the first six railcars were full and disconnected -- bumped the cars further down to access other bays. In the process, a hose snapped. The emergency cut-off valve for the bay failed. The only way to prevent a major catastrophe was for an operator to rush in to manually close the valve. He stopped the flow but received serious cold burns. Fortunately, the propylene vapor cloud didn't ignite. How can we ensure that nothing like this ever happens again?

Send us your comments, suggestions or solutions for this question by April 9, 2010. We'll include as many of them as possible in the May 2010 issue and all on ChemicalProcessing.com. Send visuals — a sketch is fine. E-mail us at ProcessPuzzler@putman.net or mail to Process Puzzler, Chemical Processing, 555 W. Pierce Road, Suite 301, Itasca, IL 60143. Fax: (630) 467-1120. Please include your name, title, location and company affiliation in the response.

And, of course, if you have a process problem you'd like to pose to our readers, send it along and we'll be pleased to consider it for publication.