First of Two Parts

Using the example system in Figure 1, this series will focus on the process elements found in piping systems. In this example, a process fluid is pumped from a storage tank, PX-TK-120, through an end suction pump, PX-PU-120, specified to pass 800 gallons per minute (gpm) with 202 feet of head. From the pump discharge, the 80 F process fluid travels to a heat exchanger, PX-HX-121, where the fluid is heated to 120 F. Level control PX-LCV-120 maintains the level in process vessel PX-PV-122 to 15 feet. The system boundaries are the tanks PX-TK-120 and PX-VP-122. The system contains only one circuit. Table 1 lists the physical properties of the process fluid.

equation
Figure 1. Example system consisting of the items making up the system along with displayed operating dataFigure 1. Example system consisting of the items making up the system along with displayed operating data (Graphics courtesy of the author)

Less than six months ago, a piping system model was created and validated using the installed plant instrumentation (values are shown in Figure 1). The difference between these values and the calculated results was less than 2 percent.

The piping drawing shows that the flow rate through the system is controlled to maintain the level in the process vessel PX-PV-122 at 15 feet. The system does not have an installed flow meter, so we must determine the system flow rate.

One of the easiest methods is to use a portable clamp-on ultrasonic flow meter. If operated correctly, these devices can provide accuracy of ±11 percent. During the assessment, the flow meter indicated a flow rate of 770 gpm.

In a previous column, we discovered that the flow rate through a pump can be calculated by converting the differential pressure to head. Using the pump curve, enter the value for pump head on the vertical axis and move horizontally until you intersect the pump curve. Then move down to determine the flow rate.

In this example, the differential pressure across the pump is 69.2 pounds per square inch (psi). Using a process fluid density of 48.9 pounds per cubic foot (lb/ft3), we can determine a pump head of 203.8 feet.

The manufacturer's supplied pump curve shows that, with a head of 204 feet, the flow rate through the pump is 770 gpm. The flow rate calculated through the pump correlates with the flow rate obtained with the ultrasonic flow meter.

Now that we have discussed how the model was validated with the observed values, we can troubleshoot.

Table 1. Physical properties of the process fluid used in this exampleTable 1. Physical properties of the process fluid used in this example
Table 2. Comparing as observed conditions with cavitation to validated resultsTable 2. Comparing as observed conditions with cavitation to validated results

An operator notifies the shift supervisor that pump PX-PU-120 sounds like it is cavitating. Additionally, the pump discharge pressure gauge is oscillating, another indication of possible pump cavitation. Table 2 shows the system's current operation along with the validated results.

Table 2 shows that the levels and pressures at the system boundary tanks are the same in both conditions, resulting in the same static head. The pressure at pump suction pressure PX-PI-120 is -2 pounds per square inch gauge (psig), 2.4 psi less than the validated results. The pump discharge pressure PX-PI-121 is 67 psi, 2.6 psi less than the validated results. According to Table 2, the position of PX-LCV-120 is 78 percent open, greater than the validated results.

The first step is to determine if these conditions are the cause of cavitation. Using Equation 1, we will determine the net positive suction head available (NPSHa) at the pump suction based on the pressure gauge reading at PX-PI-120.

As indicated in Figure 1, the NPSHa is 11.5 feet. The pump curve shows that the net positive suction head required (NPSHr) is 14.3 feet. As a result, the NPSHa is greater than the NPSHr, indicating pump cavitation is occurring. Because the pump is cavitating, the pump is probably not operating on its curve.

Part 2 of this series will use the data discussed here to determine the cause of cavitation and analyze other system problems. Read it here. See more articles by Ray Hardee here.