In this third in a series of articles on energy savings in pumps, Hans Vogelesang of Pump Support considers the factors that should be borne in mind when designing piping systems. Many issues should be examined to ensure that a system will operate in an energy-efficient manner, both at design capacity and at partial load.

The energy needed for any pumping process will always be consumed by the operating pump. In the first part of this series1, we saw that the power consumption of a pump is directly proportional to the pump capacity (Q in m3/h) and the differential pressure (Δp in bar), but inversely proportional to the efficiency rates of the pump, transmission and electric drive. Therefore, the first thing to pay heed to when aiming for low energy consumption pumping operations should always be a reduction in pump capacity and differential pressure. An energy-efficient pump set (i.e. the combination of the pump and drive) will furthermore require not only the pump to work highly efficiently, but also the transmission (flexible transmission, belt transmission, flexible coupling, etc.) and the motor.

System characteristic

In order to select the right pump it is essential to have a clear picture of your system’s properties. A very useful means of gaining this information is from the system characteristic, also known as the piping characteristic. This is a graph or chart showing the required differential pressure or differential head of the pump at various volumes of fl ow. The differential head consists of the sum of the geodetic differential head, the difference in pressure between the two tanks and the friction loss in the pipes. Because there will be no friction loss when there is no fl ow, the required differential head at a fl ow volume of 0 m3/h will always be the total of just the geodetic differential head and the pressure difference in the tanks.

The friction loss will only increase when the fluid is actually flowing through the pipeline. (The previous article demonstrates how to calculate friction loss2.)If the required differential head is calculated for different volumes of fl ow, then the results can be represented in graphical form as the system or piping characteristic. The required diff erential head for the pump consists of the sum of the geodetic diff erential head, the diff erence in pressure between the two tanks and the friction losses.

Pump characteristic
The pump characteristic can be useful when it comes to selecting a pump. These are graphs compiled by the pump manufacturer from test-bed data that give information about the differential head a pump will provide at a particular delivery capacity. The pump curve consists of a series of lines from which readings can be taken of the differential head, efficiency rate and energy uptake at different pump capacities.shows an example for a centrifugal pump. The pump curve generally applies for a particular speed and a particular fluid. If any of the factors – engine speed, fl uid viscosity or density – change then this will influence the pump’s performance and the graph will be rendered useless. However, the majority of pump manufacturers will generally quickly provide revised graphs on request.

A point to bear in mind
Once the piping characteristic is known it will be possible to establish the required differential head at the desired design capacity. It is important to realize that this will only apply under certain operating conditions. If, for example, the geodetic differential head is higher then a parallel curve, shifted to higher values, will apply. If a greater friction loss occurs in the system, as is the case when a filter gets dirty or a valve is closed a bit tighter, then the curve will rise more steeply.

Which pump?
In order to determine the design specifications of your future pump you should use the piping characteristic with the largest differential head,otherwise the pump will not be able to reach the required pressure at the design capacity. At the design stage, the pump specifications are generally based on the worst operating conditions, that is, those involving the largest conceivable geodetic differential head, the dirtiest filter, the highest cooling demand, the system operating at full capacity, etc.

The pump that is designed to meet these requirements will in practice be able to deliver sufficient pressure under any circumstances imaginable. However, as a result of the most extreme operating conditions being taken into account, the pump will only operate at partial load during normal operation.This situation leaves one facing a number of challenging dilemmas when choosing a pump.

The nature of the problem
Every type of centrifugal pump has a characteristic similar to that shown in Figure 3, whereby the efficiency gradually increases as the output grows, until a certain maximum is reached. If the fl ow increases beyond this point then the efficiency rate begins to decrease again (as shown by the green curve in the fi gure). There is therefore an operational point of maximum efficiency, at which the pump’s losses are at their lowest. This is the optimal operational point and is often marked on the pump graph as the BEP (best efficiency point) or ηmax.

To save energy the best pump will be one that can reach the required combination of capacity and differential head at the BEP, or as close as possible to this point. However,if the choice of pump is based on optimal efficiency rates under design conditions, whilst the pump hardly ever reaches this point under normal operating conditions,then the pump will rarely be working at maximum efficiency. Generally, a pump capacity that is too high or too low will result in unwanted extra energy loss.

When does this problem occur?

This problem of unfavorable pump capacity is encountered in many situations. that uses a central cold water pump to circulate cooling water through two parallel heat exchangers.At the design stage, the choice of pump was based on the total amount of cooling water that would be needed when all the heat exchangers were operational, but if one of the two heat exchangers is switched off then the pump will only have to deliver part of its original output. If the pump output is reduced from 100% to 50% there will be a great drop in the level of efficiency. efficiency will drop from approximately 75% at 90 m3/h to around 55% at 45 m3/h. The reduced efficiency will mean the pump consumes more energy than one whose BEP is at the required 45 m3/h.

A is the most usual arrangement for a system with several parallel circuits. The central circulation pump ensures fl ow through all the individual circuits. There is also an extra problem that is frequently encountered, in addition to the aforementioned lowered efficiency rate when the pump output is changed. The friction factors in parallel circuits are usually very variable. However, the difference in pressure taken over all the circuits together is constant (being the pressure difference between P1 and P2).

Therefore the incoming fl ow at P1 will divide dependent on the ratio of the fl ow resistance in the circuits, with the majority of the fl ow following the path of least resistance, though this will usually be more than is required. In addition, the fl ow rate in the other circuit will generally also be lower than intended. As a rule, on resumption of the process extra resistance will be introduced in this lower resistance circuit, in order to balance out the flow distribution. Control valves or modulating controlled balancing valves are generally employed for this task. It is a straightforward and frequently used approach, but it should be realized that the extra resistance will continuously be using up extra energy! The required diff erential head for the pump can easily be determined from the system or piping characteristic. The pump characteristic provides detailed information about a pump’s performance.

Avoiding these problems
how the problem of extra energy being lost through balancing can be solved. The differential

Box 1. Calculation of the pump’s differential pressure

The pump’s differential pressure is
always:
Δp = pd – ps

For a typical system the following equation holds for the suction
pressure:
ps = pts – pfs + Hsρg
where the pressures (ps, pts and pfs)
are in N/m2 (1 bar = 105 N/m2);

height Hs is in m; ρ, the fluid density, in kg/m3; and g, gravitational acceleration, is 9.81 m/s2. The discharge pressure can be calculated in a similar way:

pd = ptd + pfd + Hdρg

The total differential pressure of the pump can then be calculated using the following formula:

Δp = (ptd – pts) + (pfd + pfs) + Hgeoρg
Note, when referring to centrifugal pumps it is more common to talk of the total differential head, Ht or TDH, in metres of liquid column, rather than differential pressure. If the differential pressure needs to be converted into a comparable differential head then this can be done as follows:

Ht = Δp/ρg = [(ptd – pts) + (pfd + pfs) +
Hgeoρg]/ρg = Hgeo + (ptd – pts)/ρg + (pfd + pfs)/ρg

There are three different factors involved here:

b) The difference in pressure between the two tanks,

c) The friction loss in the suction and discharge pipes. Various designs for branched systems. Example A is the most frequently used, but consumes the most energy. Examples B and C are energy-efficient versions.

head for the main pump is mainly calculated for the circuit with the least flow resistance. As the pump is not able to produce sufficient pressure for circulation through the circuit with the highest flow resistance, an extra booster pump has been added to that particular circuit, to increase the pressure to the required level. Although this means an extra pump being employed it does ensure no energy is unnecessarily lost to balance the flows.

However, in example B there remains the problem that when one of the two heat exchangers is switched off the main pump will have a much smaller output and will consequently be much less efficient. How this problem can be solved is seen in example C, where a separate circulation pump is used for each individual circuit. Moreover, each pump will be fully geared to the capacity and required differential head for its own particular circuit. This ensures that if one of the circuits becomes non-operational then the other pump will always continue to work with maximum efficiency at its design capacity.

Conclusion
If a system is to be designed in an energy-efficient way it will be necessary in the first place to ensure the efficiency rates of the pump and drive, etc. are as high as possible. In addition, each pump must be made to operate as much as possible at its BEP. Moreover, it is important to avoid unnecessary energy being lost with balancing valves or orifices. Hence it is often worthwhile to install more pumps, even if this is not common practice. From a careful and accurate assessment of the various sorts of designs it will become apparent which type can provide the best solution in terms of energy efficiency.