- The Filtration Society - https://www.filtsoc.org -

Volume 16, Issue 2

GKD – Gebr. Kufferath AG (pages 95-97)

This paper provides a description of an alternative (disc filter) technology to conventional sand filters. A new Plain Dutch Weave mesh called ODW 6 is discussed, and some performance indicators are presented.

HazmanHasib, Allan Rennie, Neil Burns and Louise Geekie (pages 97-102)

This research utilises additive manufacturing technology to fabricate filter mesh designed with non-stochastic lattice structures. Disc filters with 1-layer, 2-layer and 3-layer thicknesses of repeated 1.8 mm lattice unit cell as the filter mesh are modelled in SolidWorks. Computational Fluid Dynamic (CFD) simulation using ANSYS CFX is performed at eight different flow rates (250-390 L/min) and the results (pressure drop and velocity) are analysed. Simulations are also done for perforated plates with circular-shaped and square-shaped holes with the same aperture size and filter cut point for benchmarking purposes. The outcomes indicate that the pressure drop of the lattice filters is noticeably lower than the perforated plates. These findings show that several layers of lattice structure could be stacked together as filter mesh to increase filtration efficiency with minimal pressure drop and to create a more tortuous path for the fluid.

Susan Guibert, Grace Handford, Jim Lennington, Tyson Decker, Michael Bourke and Stacey Bickler (pages 103-111)

This paper discusses the operation of a 4 MGD pressurized two stage ultrafiltration (UF) plant over a 14 month period at the Oliver-Mercer-North Dunn (OMND) drinking water treatment facility. The OMND treatment facility utilizes UF followed by low pressure Reverse Osmosis (RO) softening to produce a blended finished water capacity of 3.6 MGD. The 4 MGD capacity of the UF system supplies the feed water for the softening system and a softening bypass system for a blended finished water consisting of 60% softening system product water and 40% UF product water.The plant is located approx. 12 miles (20 km) northwest of Beulah, North Dakota, USA and uses Missouri River water from the Lake Sakakawea reservoir as a source of supply.

The plant experienced a minor issue during start-up in that a shipment issue resulted in all the UF membranes being subjected to sub-zero temperatures (-20ºC) during transportation to site. Unlike flat sheet polyvinylidene fluoride (PVDF) membranes which are shipped dry, pressurized hollow fibre membranes are shipped wet with a preservative solution to ensure the membranes do not dry out. Subjecting the modules to freezing temperatures can result in freezing of the preservative solution as well as drying out of the membrane fibres. Extreme cold, sub-zero, temperatures can cause the fibres to become brittle and allow them to snap easily with any jarring motion. Solutions to these problems include pinning of the broken fibres and re-wetting of the dried fibres with a 40 wt% ethanol solution. For the specific plant at OMND only 2 modules out of 150 were significantly damaged by the low temperature. Both of the damaged modules were pinned and returned to the manufacturer. One of the modules was re-wetted back at the factory. The re-wetted module recovered back to within the manufacturing Quality Control Release Value (QCRV). Five additional modules had damaged fibres but the damaged fibres were limited in number and repaired on-site. The five repaired modules, along with the other modules in the plant appeared to ‘wet-up’ during start up and became tighter, passing less air, resulting in lower pressure decay and higher Log Removal Values (LRVs) over a period of a few days.

To ensure that the integrity of all 150 modules was intact, daily manual logsheet readings along with PLC data was gathered. Data included the start pressure, end pressure and calculated LRV result from the daily Pressure Decay Test (PDT) also known as the Direct Integrity Test (DIT). To ensure a conservative result the LRV calculations used the Darcy pipe flow model. Based on the US EPA Membrane Filtration Guidance Manual1 membrane manufacturers have the option of choosing either the Darcy pipe flow model, which assumes turbulent flow in a breach in the fibre, or the Hagen-Poiseuille model which assumes laminar flow or a combination of the two. The calculated LRV values are always lower using the Darcy pipe flow model. For example, the following was written in an independent challenge test report prepared for the California Department of Public Health for a pressurized UF membrane module similar to the one used at OMND: “The removal efficiency of microspheres was measured at 4.6 and 5.3 log when the Darcy model predicted approx. 4.0 log removal and the Hagen-Poiseuille model predicted approx. 4.7 log removal.” By using the Darcy pipe model a conservative LRV result is assured.

At the OMND plant calculated LRV values are consistently 4.3 log or greater using the Darcy pipe model. Other conservative parameters were also used in the LRV calculations in determining the minimum test pressure required for detection of a 3 µm or smaller breach in the membrane surface. For the pore shape correction factor (K) the most conservative value of 1 was used. For the liquid-membrane contact angle (θ) zero degrees was used as opposed to a measured contact angle. Overall integrity performance of the plant for a 14 month period is discussed in detail. Trans-membrane pressure (TMP), flux, temperature corrected permeability (specific flux) and cleaning efficiency for both the 1st and 2nd UF stages is also presented.

Eiji Iritani, Nobuyuki Katagiri, Yukihiko Toyoda and MunetakaKamiya (pages 112-120)

Deliquoring methods have been developed for efficiently reducing the average porosity of filter cake formed by filtration operation without employing any special expression cells. In the methods, filtration was initially conducted using suspension in which particles were flocculated by controlling the solution environment appropriately to produce a high filtration rate. Once the cake formation was complete, the filter cake was further compressed by the principle of reversible flocculation brought about by permeating the solvent in which particles were maintained in a dispersed state.

In this study, the solution environment was controlled in the following three ways:
a) presence or absence of salt in the solvent
b) change of solvent pH
c) presence or absence of alcohol in the solvent.

In these methods the flocculated filter cake formed by a filtration operation in the initial stage, and subsequently the compressed cake with a relatively low moisture content was obtained by solvent permeation through the use of reversible flocculation. The suitability of the methods was determined by performing microfiltration of TiO2 suspensions and subsequent permeation in a deadend mode under the condition of constant pressure.