How does the sum in the heading make any sense? Enter a very fundamental mathematical concept, the exponent. We also need to bring in some bright minds from decades and centuries gone by. The list is long and many of the names are familiar but the sum of all their efforts confirm that losses in a piping system vary, approximately, as the square of the velocity of the fluid moving through the pipe.
Throttling pumps is a counterproductive process as it implies spending money by adding energy to a fluid and then converting a portion of this energy to heat (to no useful purpose) using an expensive valve. Very often, the valve is not suitable for this purpose and the result is high levels of cavitation and bad control characteristics.
From the diagram it becomes immediately apparent that this station was designed for a two-operating-one-standby configuration. For various reasons, they were operated virtually continuously in an all-three-operating manner. Even in a new condition, the duty point would move far to the left of each of the three pump's curve i.e. 59m.
...an oversized pump will tend to operate on the right hand side of the performance curve at lower efficiencies and with a higher NPSH required. Cost per cubic meter of liquid pumped (specific energy) and, in all likelihood, pump maintenance costs will increase significantly and plant reliability will decrease.
How many pumping systems are designed on a single duty point? If many of the authoritative statistics that come out of the market are to be believed, a lot. One figure bandied about is that more than half of the pumps in Europe are too big for the application! This in a so called first world environment. What would it be in the less developed countries? It can only be speculated that it is considerably worse. How does this happen and where does it start?
Experience in the pump manufacturing and marketing sector saw many clients phone up and ask for a pump set to do nice round duty points such as 30l/s against a Total Dynamic Head of 50m. Without getting too smart for one's own good, the question that begs answering is why not 31,6l/s at a TDH of 49,2m. Possibly the answer lies in rounding up or down coupled with allowances made for future expansion, inexperience of junior engineers or differing views as to what the real duty is. One instance of total overshoot in sizing was a platinum mine that doubled the duty point of their slurry circuit pumps as it was planned to double production after 5 years. The result was 3 x 300kW pumps in series doing the work of 3 x 110kW pumpsets. A Life Cycle Cost exercise showed a payback period of less than a year if the larger pumps were replaced with the smaller units. To rub salt in the wound, the worldwide economic meltdown occurred exactly on the (five year) date set for the doubling of production just in time for a massive slump in demand for platinum. 5 years of inflated energy consumption, poor power factors and sky high maintenance for no gain!
What are the lessons from this and too many other case studies? Simply put, life is all about variations and it is how these variations are handled that is the key to squeezing costs down as far as they can possibly go. Pump selections are about a range of duty points. These duty points start from a maximum and progress to a minimum. The trick is to bracket the Best Efficiency Point, or BEP with these two points.
More about all this later.
Part 1: NSS practical aspects and calculation
During a recent visit to a large chemical plant, I was intrigued by the unusually loud knocking noise which was emanating from three reasonably large (350kW) cooling water end suction pumps. A non-technical colleague who was with me at the time even remarked that the noise was unusual.
Closer inspection revealed an intermittent but severe knocking noise in both the suction and discharge, to the degree that the rubber compensators were showing evidence of rapid internal pressure transients. Classic cavitation symptoms caused by operation too far left of the Best Efficiency Point (BEP) combined with a high pump suction specific speed number(NSS)!
The plant supervisor confirmed that it was non-standard practice to run all three pumps at one time as the system was designed for a two operating one standby basis. That at least explained the operation on the left hand side of the pump's curve (see last posting on system head curves).
The pumps in question were designed for a high capacity/low head type of operation where NPSH required becomes something of an issue. In order to meet the challenge, the pump designers enlarged the impeller inlet area to reduce velocity and use Bernoulli's theorem to increase the pressure in the eye of the impeller. This solved the problem of having an NPSH required which was too high, but created another problem in the form of recirculation of flow within the eye of the impeller. The resultant drop in pressure often reaches a point which is less than the liquid's vapour pressure. This allows the formation of vapour pockets which implode as they reach areas where the pressure has increased.
Getting back to the NSS, the Hydraulic institute recommends a number less than 175 (metric). Design values higher than this will increase the chance of pump failure exponentially. The formula used for calculating NSS is:
Where: Q = flow rate in cubic metres per sec at BEP
NPSHr = Net Positive suction head at the same impeller diameter and flow rate
Part 2 of this blog will cover the relationship between impeller eye area and NPSHr.
This is the first in a 2-part series. Read Part 2 here.
Over the past year the profile of delegates attending courses offered by JTA have ranged from professional engineers right through to pump operators. During the courses, many were exposed to the features and benefits of systems head curves for the first time. The level of interest shown in this particular subject has been amazing. Most graduate engineers say something like "yes we did this at university but very briefly and we were not shown where and how to apply this". When the theory associated with system curves was related to real life case studies then the light bulbs really started coming on! For a trainer to see eyes lighting up and comments such as "so THAT's why that station is giving us so much trouble!" coming out, the rewards are hard to express.
What is a system head curve? Simply put, it is a graph which shows the total dynamic head (TDH) in a pipework system at various flow rates. Often this is superimposed on the performance curve of an appropriate pump and the duty point will be at the intersection of the system curve and the pump's performance curve. If the static head varies for any reason, a second system curve can be drawn which shows the increased TDH. The two intersection points show the range of flows and heads the pump will "see" as the static head changes.
Where does this all originate from? Thanks must go to a colleague who took the time to take this trainer through the process (too) many years ago. At that time the light went on and, yes, somebody was found to be at home. The lesson stuck and whenever the suspicion arose that the system was the culprit, out came the graph paper (this was before Excel and other dedicated software) and the friction loss charts. Once the plotting process began the story unfolded before some eager eyes. Yes the pipe was too small or in some cases too big!
Once all the plotting was complete, there it was, in black and white or in colours, the whole sad story. No more arguing or debate. The solution(s) became pretty self evident and the feeling of achievement and satisfaction were palpable.
All that trouble to modify the baseplate to accommodate a overmount belt drive arrangement, the bigger motor with the significant increase in energy consumed, all for a very modest and inadequate increase in flow rate. The two possible solutions? the first was replace the existing pipe with 2500m of one with a larger diameter. The second option was to use the existing line and build a bigger sump in order to accommodate the peak flows and give the pumps time to handle the temporary peak flow. The client was, understandably reluctant to purchase new pipe, remove the existing line and install the larger pipe in it's place. I must admit that his eyes did light up when I offered the larger sump solution so this poor pump salesman never got the order but later events showed he gained a trusting client!
This is the first in a 2-part series. Read Part 2 here.