Most media selection mistakes stemmed from engineers choosing a product based on a data sheet before characterizing the water to be filtered. Sand, anthracite, garnet, and zeolite are all viable media if the engineer uses them in the suggested applications.
This guide will show the diagnostic steps to take before comparing specifications, and it will explain how the various media types are related to the problems identified with the feed water. The following steps should be taken in order:
Start with water, not a catalog
Before reviewing a catalog, engineers should determine the turbidity, total suspended solids, pH, and other relevant constituents such as iron, manganese, ammonia, and organics in the water to be treated. This seems self-evident, but it is frequently skipped because engineering managers or procurement officers already have a media type selected before water analysis begins.
This is the most crucial step in media selection because treatment and filtration are not the same things. A bed that has been proven to remove the finest particulates may not have been tested for its ability to remove dissolved ammonia. Media that has been selected for its particulate removing qualities will allow dissolved iron or manganese to pass through the bed and precipitate downstream, where it was not expected.
The water analysis serves as a reminder that different mechanisms are required to separate dissolved constituents from water.
Coagulation and flocculation pretreatment significantly impact the particle size distribution of the water entering the filter. If chemical coagulants have been added to the raw water feed to promote particle aggregation prior to filtration, particle size distribution and zeta potential have been altered, favoring large flocs that are more easily filtered.
In other words, media that would have been ineffective on raw water may work well on pre-treated water, and media that would have worked well on raw water may be ineffective on pre-treated water. Make sure that the numbers used in media selection decisions represent raw or pre-treated water, as the case may be.
Effective size dictates expansion and depth of penetration; uniformity coefficient dictates fluidization and density
The two most essential characteristics on a media specification sheet are effective size (ES) and uniformity coefficient (UC). These characteristics define the media particle size distribution, which in turn dictate the hydraulic performance of the bed and its ability to clean itself during backwash.
A media bed with an undesirably low ES packed too tightly will cause the filter to blind rapidly, resulting in excessively high head losses after only a few hours of clean filtration. Similarly, a media bed with an undesirably high UC will fluidize unevenly during backwashing, creating pockets of stagnant water that will shorten the backwash cycle and leave the bed partially dirty.
Media beds that do not backwash properly will rarely be encountered in a specification sheet but will always be encountered in a pressure gauge within the first few months of commissioning. The ability to backwash properly is critical to hydraulic performance and should be discussed with the supplier in selecting the media bed.
Specified hydraulic loading rate should match the intended use
Rapid sand filters are typically sized to handle 4 – 10 m/h of hydraulic loading rate. Multi-media beds, on the other hand, have greater flexibility because they can be designed to use a combination of media with varying densities and effective sizes. The multi-media beds are used extensively in the industry because they provide greater flexibility in hydraulic loading rates which are often dictated by the flow rates of the host facility.
The biggest mistake in specifying hydraulic loading rates is to choose a value lower than the peak demand of the facility. If the facility is expected to have increasing demands in the near future, this too should be factored into the final selection. A consistently increased loading rate beyond the rated value shortens the time between backwashes, increases the differential pressure across the bed, strains the backwash pumps, and accelerates the wear and tear of the system beyond the maintenance schedule.
Design the backwash cycle in tandem with the media selection
The backwash cycle is frequently overlooked when designing a media bed. If the bed cannot be washed effectively, the desired performance and quality of filtration will never be achieved. The media bed will become dirtied, blinded, and uneven throughout its lifetime, resulting in its inability to perform close to its design specifications. To put it another way, granular media beds should be expanded by 15 – 30 percent during backwash to clean adequately.
This information is related to effective size and density and should be obtained at the same time. The expansion rate during the backwash is related to the flow rate and pressure needed to fluidize the bed, which must be specified at the same time as the media bed is being sized. If the expansion rate is lower than the recommended value, the media will fail to clean itself after a certain number of cycles, leading to gradual channeling within the bed.
As a result, if the backwash pump has been selected before finalizing the media bed specification, it is necessary to double-check that the same pump can provide the required flow rate for the selected media bed. It is preferable to spend extra time calculating the backwash cycle to ensure that the correct pump has been chosen than to change the pump halfway through the project.
Choose the appropriate mechanism based on the contaminants to be removed
Once water analysis and hydraulic design have been completed, media selection is relatively straightforward. Choosing the appropriate mechanism to remove the targeted constituents from the water is more important than choosing between brands.
The surface mechanism of straining is the simplest and most effective method for removing the largest particulates. It is, however, not recommended for high-solids water because the differential pressure across the bed builds up quickly and necessitates frequent backwashing. The depth mechanism is used in multi-media beds to remove the finest particulates.
Adsorption is a type of filtration that removes organic matter, chlorine, and other dissolved substances from water. Activated carbon media is used in adsorption filters. Granular media does not have the ability to adsorb anything because it does not have the surface area or the right chemistry to do so. Ion exchange, on the other hand, works by removing charged impurities such as ammonia, hardness, or metals.
The majority of media failures are caused by a mismatch between the mechanism and the targeted constituent. The rest stems from an inadequate understanding of the constituent’s size in relation to the media bed’s depth.
Use lifecycle costs, not just purchase costs, to select media
The purchase price of media is not the most important factor. Water consumption during backwashing, media replacement, and electricity usage during backwashing all contribute to the overall cost of a filtration bed. If a light media that requires frequent backwashing is chosen over a heavier media that needs less backwashing, the former’s operating costs will quickly surpass the latter’s purchase price.
The same is true for media that requires frequent replacement. Over a five-year period, a media with a high purchase price but infrequent replacement intervals may prove to be the most cost-effective option. Use lifecycles to calculate costs rather than installation costs. This is especially important when selecting ion exchange media because it has a direct impact on the cost of regeneration, which is frequently overlooked in comparison to the media acquisition costs.
Consider the option of using exchange media with inherent adsorption properties
Ion exchange is a powerful mechanism that can be used to remove undesirable constituents from water. It was previously discussed that granular media is ineffective at removing charged constituents because it lacks the necessary surface chemistry. Ion exchange resins, on the other hand, have been designed to remove charged constituents, providing excellent results.
Natural zeolite, on the other hand, has some of the same characteristics. Because of its natural cation exchange properties, Australian natural zeolite is a viable option when there is a requirement to remove both particulates and dissolved ammonia, as well as other metals. In terms of performance, it is similar to sand filtration for particulates.
However, it also has the ability to remove significant amounts of ammonia, which is ideal for municipal wastewater treatment plants as well as closed-loop aquaculture and industrial water systems with similar water quality requirements.
Size a multi-media stratified bed to ensure that the density and effective size differences between media layers allow for adequate backwash expansion to reestablish stratification
Anthracite is placed on top of sand and garnet in a multi-media bed because it has the lowest density and the largest effective size. From top to bottom, layer the media in order of decreasing density and effective size. The density differences between the media layers are critical in ensuring that the media bed can be adequately fluidized during backwashing, and therefore cleansed effectively.
Without density differences, each layer of media will have approximately the same expansion rate during backwash, resulting in an inability to fluidize the bed properly. Specifying the media layer without regard to the differences in effective size and density between the layers will lead to poor stratification after a few backwash cycles, defeating the primary reason for using a multi-media bed.
Set anchor requirements based on a compliance number rather than a guess as to when the bed is “dirty enough.”
A good example is the EPA Long Term 1 Enhanced Surface Water Treatment Rule, which requires conventional filtration to remove at least 0.3 NTUs of turbidity in 95 percent of the monthly samples. This is a standard for water filtration using granular media, so it can be used as a reference. The 0.3 NTU threshold represents the lowest level of performance that can be expected from a well-run filtration system.
If the performance of the system falls below this threshold, it may indicate that the system is undersized. Designers who have carefully followed all of the recommendations in this article should not expect their systems to achieve the theoretical performance indicated in the catalog specifications. The quality of the raw water entering the system varies, the quality of the media bed is inconsistent over time, and the ability of the system to backwash itself decreases as the system ages.
Conduct a pilot-scale column test to identify problems with fouling, backwashing, and breakthrough before designing a full-scale system
The column test will reveal key information about fouling and breakthrough that a vendor’s data sheet will not contain because such tests are typically performed on synthetic water rather than the raw water that will actually be used in the system. As a result, column tests are a valuable tool for discovering problems with fouling and breakthrough.
They also help to identify problems with backwashing, which is often overlooked when designing media beds. The column test is an excellent way to spend a few weeks investigating these potentially costly issues in order to avoid them in the future. A pilot study in column form is less expensive than a complete system design that fails due to these problems.
Selecting the appropriate filters is not a catalog search, it’s a procedure that runs through a series of design steps. It begins with determining what needs to be removed from the water based on water analysis. After that, the system is sized to remove these constituents using hydraulic calculations, backwash design, mechanism choice, and so on. All these steps will eventually lead to the choice of a media bed, which may differ considerably from catalog specifications. Any media that appears similar on the surface can have vastly differing performance characteristics when it comes to removing specific constituents in a given set of circumstances.
