How do you design a water filter like an EXPERT?
I’ve seen poorly designed filters… Badly installed filters… and shockingly operated filters! One operator decided that an open bypass valve meant he never needed to run a backwash!!! Yes… we saw that at an inspection once… don’t worry it wasn’t a potable plant. But still… aside from better hiring, how do we get the most out of filters, adsorbers and exchange vessels? Follow this 3 part series where I take you through some key elements of 1) Design 2) Installation 3) Operation. Follow us on LinkedIn to catch the next posts. Engineers think water filtration design is simple, and it is to an extent… But getting cost efficiency and optimal performance is usually beyond even most professionals.
Key 1: Balance filtration rate and EBCT
Its not about the smallest filter you can squeeze the flow through. Or the largest exchanger you can afford. Every type of media has its own filtration rate range within which it works best. Too slow or too fast and you get “channelling” where the water creates a pathway to short circuit its way though. See below how different those rates could be:
Sand or GAC filters: 7 - 13 m3/m2/hr
Turbidex or certain crushed Glass filters: 20 - 45 m3/m2/hr
This is where you may target the higher number in the range and save dollars on a small diameter vessel. Thats ok for a pure particlate filter, but if there is any catalyst reaction (e.g. greensand oxidation) or adsorption (e.g. Activated Carbon) or Ion Exchange happenning… then you need to consider Empty Bed Contact Time (EBCT). It is defined as the theoretical time water spends in contact with the filter media (as the name suggests, the calculations are based on if the bed were empty to make things simpler). EBCT may not matter for sand/particulate filtration, but it certainly does if you are targetting adsorption, absorption, or ion exchange.
As an example, target EBCT’s for Granular Activated Carbon (GAC) could change depending on the contaminant of concern:
GAC for Chlorine removal: 5 minute* EBCT
GAC for Organics removal: 10 minute* EBCT
GAC for flouride or persticide removal: 15 minute* EBCT
GAC for PFAS removal: 20-30 minute* EBCT
*Note: these also depend on the exact media you purchase… the gold standard is always to do a lab or pilot study to prove your design factors.
All else being equal. A larger diameter vessel, with a slower filtration rate, will give you more EBCT. So which do we want? A smaller diameter with a higher velocity or a larger diameter with a slower velocity? The answer, is to configure the filters in series or parallel to achieve your both your desired filtration rate and your required EBCT.
Figure 1. Water filters in series vs parallel
As you can see doubling the filters in parallel halved our filtration rate and doubled our EBCT. Doing so in series didn’t effect our filtration rate but still doubled EBCT. Using these phenomena you can modify vessel diameter and configuration to get exactly the filtration rate sweet spot and EBCT you need.
Here is where the “EXPERT” comes in... Because in water treatment 1 + 1 = 0 or 3. The EBCT in the “parallel” and “series” scenarios are exactly the same however the filter performances will be drastically different and may make or break your compliance.
Things to consider:
In parallel whenever you backwash you will never achieve your original polished effluent results. In series your final polishing vessel remains cleaner right up until breakthrough.
In parallel you may reach breakthrough failure with 20-30% capacity left in each vessel. You will need to change the media and throw money in the bin. In series you are more able to fully exhaust the lead vessel before scheduling a replacement. You can even sample prior to the final vessel to guarantee outlet results.
Some media, for the same EBCT, requires slower filtration rates to foster adsporption and kinetics which don’t wash contaminants back into solution.
If you are looking for an Australian water treatment expert to help you work out these issues, feel free to contact us for a chat:
Key 2: Select the right backwash rate
You would have noticed that the backwash rate in the examples above stayed the same. Thats because we kept the vessel diameters the same. Backwash design is another key area where the masters stand out from the competition. These choices lead to some of your biggest CAPEX costs (valves, automation, storage tanks) and some of your biggest OPEX costs (labour, downtime, power). This topic is honestly too big to cover everything without writing a textbook. From Air scour, to gravel vs plenum floors, layer stratification and more… Here I will just cover the biggest problem of them all. Insufficient backwashing effect.
I once purchased tens of thousands of filter media from a supplier. I asked them for their backwash rate and I was told he didn’t know that number… but I’d need to use a 10KW pump. There are so many things wrong with that answer its hard to know where to start! He didn’t know what size diameter vessel I needed. The area of which (say 2 m2) should be multiplied by the backwash rate (say 35 m3/m2/hr) to give the flow: 70m3/hr. Even if he did guess the filter size. Telling me the motor power could either be a high pressure, low flow or low flow high pressure pump. I can only imagine how many poorly designed plants his media must be in.
I got a sample, did a lab scale test, determined the 35m3/m2/hr backwash rate and continued with my design. During my lab testing where I could see the media perform along with my years of field operation, the following is what I now distill into my designs:
Don’t cut it close. If the spec sheet says 35m3/m2/hr then design for 40 m3/m2/hr. Its cheaper to upsize now than later when process conditions change or your operations team swap out to a different supplier.
Allow spare expansion, it can’t hurt. The same backwash rate (35m3/m2/hr) will expand different height beds by different amounts. The bottom opf the bed will expand the least and if you don’t get it lifting then you are going towards a dead end. Do yourself a favour and select a slightly taller vessel, and a little higher backwash rate. You will even save water, power and time by a more efficient backwash!
Don’t rely on the top distributor to prevent media loss. Contrary to far too popular belief, thats not its job! Its a “distributor” not a “strainer”. It stops the incoming water jet borrowing a hole down into your media. Its promotes a nice, laminar, distributed flow down through your media. To prevent media loss: a) know your backwash rate b) allow sufficient expansion c) design your pump and flow control properly.
I’ll cover it in part 3, filter operation. But the number 1 issue I encounter is a designer that selects a correct backwash rate and then an operator that runs a lower rate due to some initial media loss during commissioning (which is preventable). Solving that issue I’ll leave with you.
Again, if you are still unsure, feel free to contact us for an obligation free chat. Follow us on LinkedIn to be notified of the next parts in this series.