Cutting Costs in Pump and Pipe Sizing

You specified your piping system and identified pumps that will do the job. The job is finished, right?

In the past, the answer would have been "yes." But a new design technology allows engineers to do much better. By bringing cost data directly into the pipe- and pump-sizing process and accessing a new intelligent piping system design technology, plants can achieve a more cost-effective design.

How much more effective? Recent applications at a major chemical company showed first cost reductions averaging 10 percent and as high as 17 percent, and life-cycle cost reductions averaging 50 percent and exceeding 70 percent in one case.1

The technology that accomplishes this is, in reality, two mature technologies that were recently combined.2 The first technology is one that is familiar to most piping system engineers ," pipe network simulation software. Computer programs to calculate pressure drop and flow distribution in pipe networks have been available for more than 30 years. Some companies still rely on in-house developed programs and spreadsheets, while others choose PC-based commercial software. No matter what the form, these modeling tools allow engineers to evaluate the performance of complex piping systems before any hardware is purchased.

The second technology ," numerical optimization ," is less familiar to piping system engineers. Numerical optimization methods take an engineer's design and change it to effect improvements such as reduced costs. Such methods have been available for more than 40 years, and have matured into standard usage in a number of industries.3,4

By combining these two technologies, engineers can create an automated and intelligent way to search for low-cost designs. This capability takes on even greater significance in today's competitive environment and its compressed project schedules and tight budgets.

Recent studies have shown that pumps consume approximately 20 percent of the world's electrical energy.5,6 This energy generation is costly to industry and impacts the environment considerably.

When used to minimize the life-cycle cost of newly designed pumping systems, the new optimization technology can significantly reduce cost and industrial energy usage.

It is often assumed that an experienced piping system engineer can use analysis to reach a good design. And this is true if the engineer defines a good design as one that just functions properly. However, if the engineer wants to attain the lowest-cost design, he or she likely will get bogged down in the billions of potential design parameter combinations that exist in even simple piping systems.

It might be helpful to distinguish between the related engineering activities of analysis and design. Engineering analysis is a process in which an engineer specifies a system and then uses software, a spreadsheet or hand calculations to evaluate the system. If the results are not acceptable, the engineer modifies and re-analyzes the system, repeating this process until an acceptable design results. To analyze a system, therefore, the engineer first must specify the system.

Engineering design, on the other hand, is the process of determining what the system should be. The output from analysis is the performance of a given system. The output of design is the system itself.

The differences can be further clarified by looking at the inputs and outputs. Table 1 compares these differences for piping system engineering. In piping system analysis, the engineer specifies as inputs the pipe sizes, pump sizes and components and equipment. The outputs consist of performance parameters such as flow rates, pressures, velocities and net positive suction head available (NPSHa). For piping system design, the inputs are the required flow rates, pressures, velocities and net positive suction head required (NPSHr). The outputs are the pipe sizes, pump sizes and fittings.

To get the lowest-cost design, chemical plants need an intelligent way to search for these designs among the billions of possibilities, thereby augmenting the expertise of the piping system engineer. Piping system optimization offers such a method.

Fig. 1 shows the logical structure of a piping system optimizer.7,8 The piping system layout and design requirements are specified in the input area of the user interface. Data for piping costs are assigned. Such data can be rough estimates (e.g., steel pipe costs X number of dollars per pound) or detailed estimates in which the costs per length for different pipe sizes are entered.

The piping system layout and design requirements are specified in the input area of the user interface.

After the user interface comes the hydraulic solver. This consists of conventional pipe network analysis algorithms. Here, the initial piping system design is solved hydraulically.

If this were an analysis, the hydraulic results ," the flow rates, pressures and velocities ," would be passed back immediately to the user as output. But something different happens here. The hydraulic results are instead passed to the optimizer, which modifies the original design and passes it back to the hydraulic solver. The hydraulic solver provides a hydraulic solution to the modified design. The hydraulic solver functions similarly to a subroutine, which the optimizer calls repeatedly after making design modifications. This allows a sequence of designs to be intelligently evaluated and compared.