Cooling Towers Installation

Process cooling selection is a skilful decision, as it depends on variety of variables like heat load, temperature differences, and extent of cooling, geographical location, resource availability, ambient conditions and cost of operation. A wise decision at inception returns for a life.

For example, in desert-like geography or average cold ambient areas, air cooling may be cheaper, whereas water cooling can be preferred, where ample water is available and lower atmospheric dew point areas prevail.

Here, we shall focus on water as cooling media, conventionally called as Cooling Water. It is a dominant utility and attracts a good amount of ‘Cost of Operation’. Improper design of cooling water system will impart Process Inefficiency in system and drain a lot of cost, always.

Cost control opportunity in cooling water system is enormous right from selection, installation, operation and maintenance. Our target of discussion in this article are-

  • Installation of cooling towers : At ground or at Highest Level floor
  • Central Distribution network vs. Discrete distribution network
  • One pump for all floors Vs one pump for each floors.

Beyond selection of type and size of cooling tower based on Thermal load and extra margins, the question comes here is to brain-storm on various aspects of Installations to eliminate inherent power inefficiencies from cooling water system, right at design stage and receive best returns for life cycle.

Installation of Cooling Towers: At Ground Vs Highest Level Floor

Industries deploy two type of cooling tower installations. One is On Ground shown in Figure-1. But installations at Highest Possible Elevation in the premise are very rare as shown in Figure-2. The second option is much better in terms of power cost saving and on other factors. Table-1 narrates distinct pros and cons of Conventional and Highest Elevation installations. A numeric sample calculation at the end will flash economical advantage of it. Careful citation on Table below speaks all about it.

Cooling Tower At Ground
Figure 1: Cooling Tower at Ground
Cooling Tower At Top Floor
Figure 2: Cooling Tower at Top Floor

Table 1

Table 11

Table-2 show power cost comparison for both the options, Considering 100 TR cooling tower, for 5 Deg Delta T range, with equipment requiring cooling water, installed at 20 mtr elevation from ground. Circuit has same equipment and distribution piping network.

Table 2

Table 22

Nozzle header of second installation option 10 mtr. above highest point in C W distribution circuit.

Savings in fan power cost per year are not accounted here, we can avail advantage on power saving, as a function of TR rating of cooling tower and pumping head difference. For big capacity cooling towers, numbers of modules of say 500 TR can be installed at top floor level, leaving flexibility margin.

It is empirical to install cooling towers at highest possible level in plant, forming monomeric type piping circuit with proper citation on their plant specific installations.

Pumping Heads
Figure-3: Pumping Heads

Piping Distribution: Central vs. Discrete Distribution Network

Cooling water, being a key and most widely used utility, proper distribution and smart instrumentation controls can return great amount of money on operation cost account with additional feature of flexibility. We can preserve on pumping cost, fanning cost and control evaporative losses as well drift losses.

Large chemical or petrochemical complexes preferentially select Central Cooling Water system from ease of operation and control point of view. This offers advantage of little remote installation of cooling towers from main plant building associated with advantage of uninterrupted air flow to towers and reducing corrosion potential to exposed assets from drift deposition. This is a good choice for mono products plants and continuous process plants. Whereas multi product and batch process plants operate preferably on demand and supply logic, with wide variations in operating parameters and process sequencing. Process or utility or both contribute to widely distributed pattern in Peak, minimum and maximum demand of heat loads. Central utilities particularly cooling water system are not the ‘Best choice’ for such operations. Refer Figure – 4 and 5.

Central Distribution
Figure- 4: Central Distribution
Discrete Distribution
Figure-5: Discrete Distribution

Central distribution affects flexibility. We shall be operating whole cooling water system including tower, fans, pumping and piping system, unnecessarily, if one or two sets of products are to be manufactured in same or different building set ups. Such operations attract extra power and maintenance cost. Instead, discrete cooling towers for different services will-

  • Impart freedom of selective circuit operation based on market demand contributing to controlled OPEX allocation.
  • Smaller diameter Pipe sizes required compared to central distribution case, saving capital investment on piping, accessories and structural costs.
  • Phase wise installation of just required size, saves a lot on Capital cost and interest cost.
  • Smaller cooling towers adhere lesser capital cost on equipment and civil cost.
  • Well controlled flexibility of fan and pumping operation costs. A big pump and fan are not operated for part load.
  • Spares inventory cost will be much less as is distributed between many identical installations.
  • Lesser tooling and manpower cost in case of discrete cooling tower installations.
  • Variation in thermal load of one product, do not affect the other products and utilities. It is a lead factor contributing to life, efficiencies, process safeties and product qualities.
  • Plant and Costing managers have a good control on total OPEX.
  • Disadvantages are extra space requirement, wide pipe racks, increased operators and cluster of piping distribution and its maintenance.

It is an opportunity to process and design engineers to inherit efficiencies at inception stage.

Distributed Network
Figure-6: Distributed Network

One pump for whole circuit Vs. One pump for each floor / cluster

Most engineers design cooling towers with one operative and one stand by pump for whole circuit, irrespective of central or discrete installations.

This is an ‘Inefficiency at Inspection’ stage that Drains profits for entire life cycle.

Combination of discrete and cluster distribution of cooling water connected with one or more towers will deal enormous savings on operation and maintenance cost.

It is a designer’s choice and management decision.

In most process industries, equipment operating on cooling water are distributed floor wise and in clusters. Pumps are selected based on total water flow requirement and highest elevation of piping header and equipment, influence pumping head. Refer figure-1.

If scientific approach is adopted to work out thermal load calculations depending on floor or clusters and pump capacities are selected accordingly will facilitate selection of pumping heads accordingly. Pump heads will be a function of highest elevation in a particular circuit, not a whole plant. This will allow engineer to select small pumps sets, just required for an assigned duty.

It is not necessary to have dedicated cooling towers for each floor. One tower with a common basin, suction header, connected with number of pumps and dedicated piping circuits can be integrated with advanced instrumentation.

Particularly in batch operations, there are opportunities to shut off cooling water circulation, floor or cluster wise. Varying demand, time cycle, peak loads, stand by parameters, if are managed manually, will lead to an erroneous operation. On the other hand, little of smart instrumentation will yield a lot on Operation Cost. The question is: What to do and How?

Table 3

Table 4

Referring Figure – 1 and 7, will make it very simple to understand the concept.

First case in Figure -1 depicts, one cooling water circuit for entire process or plant building. As described above, it has advantage of ease of operation and less complex piping network, but greater operation inefficiencies.

Figure -6, is more scientific and new age approach on distributed network based on demand and supply, which inherit selective operation of system Automatically and returns numerous financial gains.

Example below, elaborates distinction between these two aspects.

Say, a plant manufacturing XYZ product. Equipment are laid on floor 1,2,3 and 4 and connected to a common pumping system with required piping distribution network. Each floor height is 5 mtr. Total thermal load of 15,00,000 Kcal/hr at. Cooling tower is selected for 5 Deg C delta T cooling capacity. 500 TR Cooling tower is selected equipped with 300 CuMH pump.

Table-3: hereunder elaborate power consumption at 24 x 7, 330 days/year and Rs 7 per KWH power cost, Pump has to operate continuously, irrespective of thermal load demand, as one or other process unit or step demands cooling. Centrifugal pump operating at 70% efficiency.

Variation in fluctuations of cooling water demand, depending on customer equipment operation cycle, associated variation in water flow rate, line velocities and head loss. A common pump is pumping specified almost same quantity of water and head, is an assumption here.

Table -4: elaborates floor wise distributed network, where additional investment is in separated vertical supply and return headers and pipe rack, which is not a great investment. Horizontal headers at each floor, their sizes, and elevations are same as in case above.

Total 6 pumps are connected with a common suction header connected with the same cooling tower selected based on full range flow and pressure head calculation, as required at respective floor. Pressure Transmitter on supply header of each floor will signal to operate specific pump with respect to set point. Installation of VFD on pumps will further facilitate flexibility to control flow rate depending on pressure set points. Pump or VFD can further be tuned to differential temperature between supply and return headers. Higher delta will signal to speed up the pump and visa Vis. It can advanced to stop idle pumping as a function of time and delta T.  Sixth pump is selected of highest capacity and head amongst 5, as a common stand by.

Saving potential based on above are complex to calculate as an example, but are really contributing on average operation hours as assumed in Table below. It can be varied with product, process, distribution network, thermal load demand. One may find an adoptive idea for their specific requirement and deploy accordingly.

Energy optimization is a continuous improvement process and can be adopted at any stage of ‘Plant Life cycle’. Specific approach and spectrum of options, instrumentation, controls and selection of right set up can definitely lead to optimum costs.


Jalendu Vadilal Shah
Jalendu Vadilal Shah,
Dy. G M – Projects,
Chemical Process Plants.

Leave a Reply