Get a Free Quote

Our representative will contact you soon.
Email
Mobile/WhatsApp
Name
Company Name
Message
0/1000

Blog

 >  Blog

News

The Bag Filter That Looked Fine

Time : 2026-11-17

The call came on a Tuesday morning. A line that had been running at 3 mg for eight months started producing 6, then 9, then 11 on successive samples, with nothing changed in the process and a pressure drop across the filter that had not moved in a week.

The first instinct on a filtration problem is to change the element. They had changed it on Monday, and the reading had improved for one shift before climbing again.

Day One: Everything Looked Correct

Tuesday was spent confirming that the obvious things were fine, and they were.

Bath concentration was in range. Temperature was in range. The spray pressure at the manifold read 1.2 MPa, which was the commissioning value. The filter housing held a bag that had been in service four days, and the differential gauge across it read 0.04 MPa against a change-out threshold of 0.15 MPa.

That last number is the one that closed the investigation prematurely. In a bag filter system, pressure drop is treated as the truth about loading: low drop means the element is passing fluid freely, which means it is not loaded. The gauge was doing exactly what it was installed to do, and it was not answering the question anyone actually had.

Day Two: The Bag Came Out

On Wednesday we cut a used bag filter element open along its length and laid it flat.

The inside surface told the story immediately. Most of the bag was lightly discoloured, as expected. One band, roughly a hand's width, running from the top ring down about two thirds of the length, was loaded with a compacted layer of dark material. The rest of the surface had almost nothing on it.

That pattern is not normal loading. A bag filter in a correctly arranged housing loads from the outside in, and the deposit builds evenly because the flow is distributed around the circumference. A band of heavy loading with clean areas beside it means the flow was not distributed: it was entering through a narrow region and passing through a narrow region, and the rest of the surface was doing very little.

The mechanism has a name in bag filter work, and it is not partial loading. It is channelling. Once a local area loads, the flow finds the path of least resistance through the adjacent clean area, which loads faster because it is carrying more flow, which concentrates the flow further. The bag ends up with one clogged band, one washed-out channel, and a pressure drop that stays low because most of the surface is bypassed.

Day Three: Why It Happened Here

The cause was upstream of the bag and had been there since commissioning.

The inlet pipe entered the housing tangentially and pointed directly at one side of the bag support cage. Fluid from that pipe hit the same area of the bag every cycle, at the same velocity, carrying the same debris. The band that was loaded in the cut-open bag was exactly the area the inlet was aimed at.

A bag filter housing is supposed to distribute flow around the element. This one did not, because the inlet was aimed rather than diffused and there was no deflector. It had worked for eight months because the loading rate had been low enough that the band could absorb it without the rest of the surface becoming irrelevant. When the machining upstream changed a tool and the chip load rose, the band saturated in days, and the bag filter system started passing what it should have caught.

What Changed

Three things, in order of cost:

  • A deflector plate was fitted opposite the inlet, so the jet no longer aims at one part of the element
  • The change-out rule moved from pressure drop alone to pressure drop plus a fixed interval, because on this housing the gauge cannot see channelling
  • A monthly cut-open inspection of one used element was added, which costs one bag and ten minutes

The readings came back to 3 mg within two shifts and stayed there. The gauge still reads 0.04 MPa most of the time, which is now understood as a characteristic of this housing rather than a statement about the element.

What This Says About Bag Filters As A Control

A differential gauge measures the resistance of the whole element, and channelling is precisely the condition in which a small area carries most of the flow. The resistance stays low because most of the area is not doing work. The instrument is not faulty; it is measuring a quantity that stops being informative when the element fails unevenly.

The practical consequence is that a bag filter system needs a second signal. On this line it is a fixed interval. On others it is a particle count downstream of the element, or a turbidity reading, or a scheduled inspection. Any of those would have caught this in week one.

What I Still Do Not Know

I cannot tell you how often to cut a bag open in a housing that has already been corrected. Once a month has been enough for two years, but that is one installation, and the number has no basis beyond experience.

What I would like is a way to detect partial loading in service, without opening the housing. A downstream particle count would do it, and costs more than the bags. Whether that trade is worth making probably depends on what a contaminated batch costs, which is a figure most plants only learn after the first one.

Get a Free Quote

Our representative will contact you soon.
Email
Mobile/WhatsApp
Name
Company Name
Message
0/1000