Pump System Improvement

Editor’s note: For a more detailed analysis of the application and processes described in this article, check out Part 2 by clicking here.

I recently received a comment from a reader who pointed out that all examples in my Pumps & Systems columns are small process systems with only one or two circuits. She asked if the same process could be applied to a larger system with multiple cooling loads in parallel. Looking back, most of the examples in my columns were selected to limit their complexity so I could concentrate on one aspect of the system at a time. This month, I will demonstrate how the same approach can be applied to a system of any size.

Regardless of the size, arrangement or design objective, all piping systems are made of three primary elements. The pump elements add all the energy to move the fluid through the system, the process elements use that fluid energy to make the product or provide the service, and the control system improves the quality of the product. As we have seen, the head developed by all the pumps is equal to the head loss in the process elements and control elements in every circuit of the system.

Most industrial plants have extremely large cooling water systems that provide water to hundreds of cooling loads. Balancing the energy in larger systems is tedious to calculate by hand, but computers can quickly handle these calculations. Software can be used to build piping system models and calculate the balanced flow rates and pressures, providing a clear picture of how a system works and what can be done to improve the system.

This real-world example shows how a system energy balance was used in a large cooling water system in an industrial plant using such software.

The System & Challenge

During a software training course, an attendee presented a project involving his plant’s cooling water system. The system normally operated with two cooling water pumps providing sufficient cooling water to meet all plant loads. On a hot summer afternoon, one of the two operating cooling water pumps tripped. Staff immediately started the standby pump to return the system to normal operation before notifying the maintenance group.

When electricians inspected the motor on the tripped pump, everything appeared normal. They notified operations after checking the motor loads on the two operating pumps and discovering their motors were operating well into their service factor. After a short discussion, operations personnel started the third pump and modified the plant’s operating procedure to operate all three cooling water pumps. With all three operating, the motor loads returned to satisfactory conditions.

The plant’s operating guidelines state that for critical systems, the failure of one piece of equipment cannot cause the system to fail. The cooling water system was critical for continued plant operation, and with all three installed pumps operating, personnel took immediate action to purchase and install a fourth pump to provide the necessary standby protection.

The plant engineer wanted to see if piping simulation software could determine if the changes would correct the problem. The engineer took steps to determine the cause and what could be done to return the system to normal operation as quickly as possible.

The first step determined how the system would operate with three pumps. The plant cooling water system was needed to operate the plant, so operation of the physical system could not be modified. The plant engineer decided to use piping system simulation software to discover why the system required three pumps to operate instead of the initial two-pump operation.

The System Model

An accurate model of the piping system was created using all available design data, manufacturer’s test and operating data. The flow diagrams and the piping and instrumentation diagram identified all major equipment and interconnected piping. The flow rate, head, efficiency and NPSHR as shown on the manufacturer’s supplied pump curve were entered for each pump in the circulating water system.

The data for the system’s process elements were entered. Using the manufacturer’s test data for the installed heat exchangers, the team entered the head loss versus flow rate in the system model. Next, piping details such as pipe material, schedule, diameter, along with the valves and fittings, were entered. The density, viscosity and vapor pressure of the system process fluid were specified. This was a closed-loop recirculating system, so elevation and pressure at the surge tank were entered.

Figure 1. An example of the cooling water system described in this columnFigure 1. An example of the cooling water system described in this column (Courtesy of the author)

Next, the control elements in each circuit were entered. Each load was sized for a specific rate of heat transfer. The rate heat needed to be removed from the process being serviced. The system differential temperature between the hot and cold fluid determined the cooling water flow rate through the circuits. Using this information, one can determine the differential pressure across each control circuit. Figure 1 shows an example of the cooling water system.

Testing the Model

After completing the cooling water system hydraulic model, the plant engineer ran the calculations to show how the system was designed to operate with only two cooling water pumps.

The physical piping system required three cooling water pumps to avoid overloading the pump motors, so the physical system was not operating according to the computer simulation.

Now the engineer needed the differences between the physical piping system and the simulation. A quick look at the throttle values in the physical piping system pointed out the problem. Reviewing the manual throttle valves’ position in the physical system revealed that the majority were fully open instead of set to an intermediate position to limit the flow rate through each circuit.

The plant engineer discovered a year earlier that many new loads were added to the cooling water system. After those system changes were made, however, the system was not rebalanced as had been done when the system was first placed in operation.

The cooling water system was placed in operation, and the plant operators manually adjusted the valve positions to each throttle to maintain the required outlet temperature. As each operator independently adjusted a throttle valve, the flow rate through each circuit was affected. Before long, the flow rate through each load in the cooling water system was much higher than required by the design.

As the summer progressed and the outlet temperature increased, the operators continued to open the throttle valve, further increasing the cooling water flow rate. The flow rate through the system was greater than the two pumps could handle, causing one to trip out on high motor overload.

The plant engineer now had a good understanding of how the throttle became open, but he still did not know if his model accurately represented the physical system. The plant engineer changed the piping system model to reflect how the physical system was working. After these calculations were performed, the pump suction and discharge pressure values accurately matched the pressure readings found in the plant. Comparing the motor current on the three running pumps, the plant engineer determined that the flow rate through each pump accurately matched the model’s calculated values. The engineer now had an accurate model of the total system.

Using the Accurate Model

After validating the piping system model, the engineer set the design flow rate through each circuit in the model to the design value. Using the valve manufacturer’s supplied flow coefficient (Cv) data versus the valve position for each throttle valve, the simulation calculated the valve position needed to meet the design flow rate through each circuit.

The result was a calculated valve position for each throttle valve to balance the entire system. The physical plant had three cooling water pumps in operation and all throttle valves open, so the plant engineer wanted to balance the system without affecting the plant’s operation.

The engineer presented the results of the actual system and the model to plant management along with a process to return the system to its normal operating conditions without adversely affecting plant operations. The installation of the fourth cooling water pump was scheduled to take about four months, so management decided to attempt to balance the system.

The plan involved setting the position of the throttle valves to the value calculated by the plant simulation with two pumps operating in a planned shutdown.

Balancing the System

During the plant shutdown, the operators set the physical throttle valves in each circuit to the calculated position for the two-pump operation. Then the cooling water system was started with two pumps in operation. The operators recorded the suction and discharge pressures for the two pumps.

The observed differential pressure across the two operating pumps matched the calculated results of the system model, confirming the flow rate. Another test confirmed the flow rate through each pump by comparing the motor’s required electrical power to the manufacturer’s pump curve. To confirm flow rates through each circuit, an ultrasonic flow meter was used to show whether the flow rates matched the flow rate entered in the piping system model.

As a result of these tests, the plant engineer determined that the system was balanced for the two-pump operation and was ready to return the cooling water system to its normal two-pump operation.

When the plant was brought back to service, personnel performed further checks to ensure that the cooling water system was operating properly by checking the outlet temperature of each circuit load. The plant engineer determined that all but two of the heat exchangers were operating properly.

The cooling water flow rates through these two circuits on the physical system were increased so the heat exchanger outlet temperatures were within normal operation. Once the outlet temperatures were returned to normal, the positions of the two control valves were measured.

The new positions of the throttle valves were entered into the piping system model. The calculated results indicated that the increased flow rates through each load in the system were still within the acceptable range of operation.


This example shows how adding cooling loads to an existing cooling water system results in unintended consequences for a critical system in an operation plant.

Not rebalancing the flow rates after changes to the cooling water system was the first problem. The second was that operators increased the flow rate through the plant loads without understanding how changing one flow rate affects the flow rate in the other circuits.

The final problem was that the only option the plant considered was adding a third operating pump and a backup pump. Rebalancing the system solved the problem at hand, without the need to purchase and install a fourth pump that would take four months to install, leaving the plant without a backup cooling water pump.

By developing a model of the cooling water system’s operation, the engineer could understand how the system was designed to operate. He was able to determine the source of the problem—all throttle valves open—by comparing the calculated results to the physical piping system.

Validating the piping system model with the operation of the physical system allowed him to try alternative solutions to the problem. By simulating the system with the balanced flow rate set in the loads and two pumps operating, he was able to calculate the throttle valve position needed to balance the system.

In next month’s column, with the aid of the piping system model, we will see how the plant was able to reduce system operation costs while increasing system capacity.

To follow the process described in this example, visit www.pumpsandsystems.com/large-system-optimization.

See more Pump System Improvement articles by Ray Hardee here.