Cloth doing a job

A brine cell reads a filter cloth's blinded length

A filter cloth packed with cake resists like a layer of that cake whose thickness is an integral over the shape of its holes, and a clean cloth soaked in brine between two electrodes conducts through the same holes by the same integral. So a monofilament cloth reads its own blinded length off a meter before it has seen a slurry. A spun cloth reads a tenth to a quarter short, because current uses the pores inside its yarns that a packed flow does not, and shortest where its waist is narrowest. It still ranks six cloths exactly as their blinded duty does, where a permeameter gets the order nearly backwards. And a used cloth read in the same cell reports how far it has blinded.

Worth reading first: A blinded filter cloth is a length of its own cake · A hole is a channel, not an opening · A filter is rated by the hole it does not show.

A blinded filter cloth is a length of its own cake. Once a slurry has packed a cloth’s holes, the cloth resists as much as a layer of that same cake, and the layer’s thickness is

L=∫dzφ(z),L = \int \frac{dz}{\varphi(z)},

where φ(z)\varphi(z) is the share of the cell left open at a depth zz through the cloth. The particle drops out, and what is left is a property of the shape of the holes. For a 40 tex filter cloth half a millimetre thick, LL is 0.86 millimetres. It ranked six cloths differently from their clean permeability, and it is the number that decides what a blinded cloth costs a filter press.

It is also a number nobody measures, because it is defined after the fact. A cloth’s datasheet quotes its clean air permeability, and the length only shows itself in a filtration run as an intercept on Ruth’s line, with every uncertainty of a run attached. That essay ended by noticing that the integral is not peculiar to flow. Electric current through a channel full of a conducting liquid adds up slice by slice in exactly the same way, so a clean cloth soaked in brine between two electrodes ought to read LL on a meter.

This essay works the measurement through. For a monofilament cloth the reading is exact, to the last digit. For a spun cloth it is short, by an amount the arithmetic gives, and it is short in a way that does not spoil the one thing a buyer needs from it.

Where through its thickness a filter cloth's brine reading is spent. The filter cloth, 0.48 mm thick, between two electrodes in brine: at each depth, the length of bare brine a slice of the cloth is worth, one over the share of the slice that conducts. A monofilament conducts only through its hole, 1/φ, which rises to 3.44 at the waist where 29 per cent of the cell is open; its area is 0.86 mm, the length of cake the cloth's holes are worth when packed. A spun cloth's yarns conduct as well, at 0.25 of the brine's conductivity, which caps the peak and lowers the area to 0.67 mm, 22 per cent short. The picture takes the current straight through each slice, with no sideways spreading.
Fig. 1 The filter cloth between two electrodes in brine: at each depth, the length of bare brine that a slice of cloth is worth, one over the share of the slice that conducts. A monofilament conducts only through its hole, peaking at 3.4 at the waist; its area is 0.86 mm, the packed cake length. A spun cloth’s yarns conduct too, which caps the peak and leaves an area of 0.67 mm.

The same integral, read by a meter

A slice of cloth dzdz thick, soaked in brine, conducts through whatever of it is liquid. In a monofilament cloth the threads are solid polymer and carry nothing, so the conducting part is the hole, a share φ(z)\varphi(z) of the cell. The slice therefore conducts like a slice of bare brine φ(z)\varphi(z) as wide, and its resistance, per unit of cell area, is the brine’s resistivity times dz/φ(z)dz/\varphi(z). Add the slices:

R Aρbrine=∫dzφ(z)=L.\frac{R\,A}{\rho_{\text{brine}}} = \int \frac{dz}{\varphi(z)} = L.

That is the blinded length, exactly. The meter reads the cloth as a thickness of brine, and a packed filter reads the same cloth as the same thickness of cake. The two physical processes share nothing but the arithmetic. Darcy flow through a packed slice resists in proportion to the slice’s thickness over the cake’s permeability and its open share; current through a liquid-filled slice resists in proportion to the slice’s thickness over the liquid’s conductivity and its open share. Divide out the material constant and both leave the hole.

Measured that way, the length needs a cell, a conductivity standard for the brine and nothing else. No slurry, no filtration run, no fitted intercept. It even shows where the length comes from. The figure above is the brine-length each slice is worth, and its area is the reading. Near the faces the hole is as wide as its cell and each slice is worth its own thickness; through the middle the open share falls to 29 per cent and each slice is worth three and a half times its thickness. The waist is where the reading lives, as it was where the blinded resistance lived.

A spun cloth’s yarns conduct

A spun yarn is not solid. It is sixty per cent fibre and forty per cent pore, and in brine those pores are full of brine. That is a second conducting path in every slice, through the part of the cell the threads occupy.

How well a bed of fibres conducts is a question petroleum engineers answered for rock in 1942, and the answer is Archie’s law: a porous solid full of a conducting liquid conducts at the liquid’s conductivity times its porosity raised to an exponent mm. For a bed of spheres theory gives 1.5. For cylinders lying across the current, as a yarn’s fibres lie across a current running through the cloth’s thickness, it is nearer 2. A yarn at forty per cent porosity therefore conducts at between a sixth and a quarter of the brine around it:

Le=∫dzφ+(1−φ) εm.L_e = \int \frac{dz}{\varphi + (1-\varphi)\,\varepsilon^m}.

That is shorter than LL, and the shortening is largest exactly where LL is largest. At the faces, where the hole fills the cell, the yarn adds nothing. At the waist, where seventy per cent of the slice is thread, the yarn’s path adds more than half again to what the slice conducts. The spun filter cloth reads 0.67 millimetres with an exponent of 1.5 and 0.72 with 2, against a blinded length of 0.86: 22 and 16 per cent short.

The length a brine cell reads for six cloths, as monofilament and as spun. For six named cloths in plain weave, the length a conductivity cell reads: as a monofilament, which is exactly the length of cake its packed holes are worth, and as a spun cloth whose yarns conduct with Archie exponents of 2 and 1.5. voile 0.35 mm, spun 0.33 to 0.32; muslin 0.52 mm, spun 0.47 to 0.44; poplin 0.51 mm, spun 0.45 to 0.42; sheeting 0.75 mm, spun 0.60 to 0.55; duck 1.02 mm, spun 0.87 to 0.81; filter 0.86 mm, spun 0.72 to 0.67. The spun readings run 7 to 26 per cent short, most on the sheeting, whose waist is narrowest. The chart cannot show a real yarn's spread of packing.
Fig. 2 For six named cloths in plain weave, the length a cell reads as a monofilament, which is the blinded length, and as a spun cloth with Archie exponents of 2 and 1.5. The voile reads 7 to 10 per cent short and the sheeting 19 to 26; the duck, the filter cloth, the poplin and the muslin lie between.

Flow ignores a pore that current cannot

The difference between the two processes is the reason for the gap, and it is worth seeing why one process ignores the yarns and the other does not.

Flow weights a pore by the square of its size. Kozeny and Carman’s permeability goes as the square of the pore’s scale, so a bed of fibres fifteen micrometres across has a permeability hundreds of times smaller than a cake of coarse particles, and a packed flow passes through the cake in the hole and barely notices the yarn. A cloth stops having holes before it stops passing air because of exactly that square: the yarn path is a floor that matters only when the holes are nearly gone.

Conduction weights a pore by its area and nothing else. A pore ten micrometres across conducts per unit area exactly as well as one a millimetre across, because the current density in a liquid does not care how wide the liquid is. So the yarn’s forty per cent of pore, which a flow ignores, is worth a quarter of a free path to a meter.

That is why the spun reading is short, and also why it is short by a definite, computable amount. A third weighting reads the same pores differently again: a centrifuge drains them by their entry pressure, one over their size, which is how five spins read the film a drained channel keeps. Each instrument sees a cloth’s two pore systems in its own proportion. The cell sees the cloth through a weighting that is too generous to the yarns, and how generous depends only on the yarn’s porosity and Archie’s exponent, not on any particle or slurry.

Shortest where the waist is narrowest

Because the yarn path matters most where the hole is smallest, the cloths with the narrowest waists lose the most of their reading.

How far short a spun cloth reads, against how open its waist is. For six named cloths, the share by which a brine cell's reading of the spun cloth falls short of its blinded length, against the open share of the cell at the hole's waist, for Archie exponents of 1.5 and 2: voile, waist 49 per cent open, 7 to 10 per cent short; muslin, waist 38 per cent open, 11 to 16 per cent short; poplin, waist 34 per cent open, 13 to 18 per cent short; sheeting, waist 25 per cent open, 19 to 26 per cent short; duck, waist 30 per cent open, 15 to 21 per cent short; filter, waist 29 per cent open, 16 to 22 per cent short. A narrower waist puts more of the cell's reading where the yarn's path matters most. The picture assumes every yarn packed at six tenths.
Fig. 3 The share by which each spun cloth’s reading falls short of its blinded length, against how open the cell is at the hole’s waist, for Archie exponents of 1.5 (filled) and 2 (open). The voile, half open at its waist, reads 7 to 10 per cent short; the sheeting, a quarter open, 19 to 26.

The voile’s waist is 49 per cent open and its reading is 7 to 10 per cent short. The sheeting’s waist is 25 per cent open and its reading is 19 to 26 per cent short. The other four fall in order between them, the duck and the filter cloth, both near 30 per cent, losing 15 to 22.

The shortfall is close to a function of the waist alone, since the waist is where most of each reading accumulates. That makes the correction from a spun reading to a blinded length a matter of knowing two things: the yarn’s packing and the waist’s open share. A hole is a channel, not an opening computes the second from the cloth’s construction, and the first is a yarn property a spinner quotes. Neither needs a filtration test.

The ranking survives

What a buyer of filter cloth wants from a quick measurement is not usually the absolute length. It is the order: which of several candidate cloths will cost least when blinded.

Six cloths ranked by a permeameter, by a brine cell and by the length a cake makes of them. Six named cloths ranked from least resistance to most three ways. A permeameter, clean: voile, muslin, duck, filter, poplin, sheeting. A brine conductivity cell on the spun cloths: voile, poplin, muslin, sheeting, filter, duck. The length of cake their packed holes are worth, which is what a blinded cloth costs a filter: voile, poplin, muslin, sheeting, filter, duck. The cell agrees with the blinded ranking exactly, a Spearman correlation of 1.00; the permeameter manages 0.31. The picture ranks; it cannot show how close the poplin and the muslin are, 0.515 and 0.524 mm packed.
Fig. 4 Six cloths ranked from least resistance to most, by a clean permeameter, by a brine cell on the spun cloths, and by the length of cake their packed holes are worth. The cell’s ranking and the blinded ranking are identical. The permeameter puts the duck third and the poplin fifth; blinded, they are sixth and second.

The brine cell ranks the six spun cloths exactly as their blinded lengths do: voile, poplin, muslin, sheeting, filter cloth, duck. A Spearman correlation of 1.00. The shortfall is large but it moves every cloth in the same direction and by amounts that follow the waist, and the waist was already part of what ranked them.

A clean permeameter, which is what a datasheet reports, manages a correlation of 0.31. It puts the duck third, because a thick cloth with wide holes passes air easily, and the duck is the worst of the six once blinded, because it is thick. It puts the poplin fifth, because its holes are small, and the poplin is second best blinded, because it is thin. The fourth power is a close cloth’s rule is why: clean flow charges the narrowest channel heavily and a blinded cloth charges thickness over open share.

One pair is close, and a real measurement would need care with it. The poplin’s blinded length is 0.515 millimetres and the muslin’s 0.524, two per cent apart. Spun, the cell reads them 0.42 and 0.44. A cell that could not resolve a few per cent would see them as tied, which is the honest answer.

A looser yarn steals more of the reading

The yarn’s packing enters the reading through εm\varepsilon^m, and packing varies by spinning system: an open-end yarn is looser than a ring yarn, and a compact yarn tighter.

What a brine cell reads of a spun filter cloth, against how tightly its yarn is packed. The length a conductivity cell reads for the filter cloth spun from yarns packed from 45 to 75 per cent fibre, with Archie exponents of 1.5 and 2, against its blinded length of 0.86 mm: m = 1.5, 0.62 mm at a packing of 0.5, 0.67 at 0.6 and 0.72 at 0.7; m = 2, 0.67 mm at a packing of 0.5, 0.72 at 0.6 and 0.77 at 0.7. A looser yarn conducts more and reads further short; sealing the yarns, or weaving monofilament, returns the blinded length. The picture holds the cloth's holes fixed as the yarn's packing changes.
Fig. 5 The length a cell reads of the spun filter cloth, against the fibre share of its yarns’ section, with Archie exponents of 1.5 and 2, its holes held as they are. At a packing of 0.5 it reads 0.62 to 0.67 millimetres; at 0.7, 0.72 to 0.77. Sealing the yarns returns the blinded length of 0.86, ruled in grey.

From a packing of 0.5 to 0.7 the reading rises by about a sixth, from 0.62 to 0.72 millimetres at an exponent of 1.5. So two cloths with identical holes, one woven from open-end yarn and one from compact yarn, would read differently in brine while blinding identically. The cell has a systematic error in the yarn’s packing that the blinded duty does not.

The remedy is the one the earlier essay named. Seal the yarns — a wax, a hydrophobic finish that keeps the brine out of the fibre bed, or simply a monofilament — and the second path closes and the reading is LL again. A buyer comparing spun cloths from different spinning systems should seal them; one comparing cloths of one yarn need not.

A used cloth reads how far it has blinded

There is a second use for the cell, and it is the more practical one. Put a cloth from a working filter press into the same cell, with its holes packed with cake.

Packed cake is itself a porous bed, and it conducts at its own εm\varepsilon^m. At forty per cent voids and an exponent of 1.5 that is 0.253 of the brine. So a slice of packed hole conducts four times worse than the same slice open, and the reading rises as the cake packs in.

What a brine cell reads of a filter cloth as a cake packs into it. The filter cloth in a conductivity cell as cake of forty per cent voids packs its holes from the face down, against the depth packed. Packed cake conducts at 0.253 of the brine, so the reading rises from 0.86 mm clean to 3.38 packed through as a monofilament, and from 0.67 to 2.64 spun with its yarns blinded too. Half way in, the monofilament's excess over its clean reading, divided by one over 0.253 less one, is 0.431 mm, which is exactly the length of cake its packed holes are worth at that depth. The picture assumes the cake packs at its own voidage and that the brine reaches every pore.
Fig. 6 The filter cloth in a cell as cake packs its holes from the face down, against the depth packed. As a monofilament the reading rises from 0.86 mm clean to 3.38 packed through, and spun with its yarns blinded too from 0.67 to 2.64. The steepest rise is through the waist, where each millimetre packed is worth most.

The rise has a simple form. Every slice the cake reaches has its conductance multiplied by εm\varepsilon^m, so its contribution to the reading is multiplied by ε−m\varepsilon^{-m}. The excess over the clean reading is (ε−m−1)(\varepsilon^{-m} - 1) times the cake length packed so far, the same integral up to the depth the cake has reached:

Rused−Rclean=(ε−m−1)Lpacked.R_{\text{used}} - R_{\text{clean}} = \left(\varepsilon^{-m}-1\right) L_{\text{packed}}.

For the filter cloth packed to its waist, 0.24 millimetres down, the excess reading is 1.28 millimetres and the constant is 2.95, so the inferred blinded length is 0.43, which is exactly the length of cake the cloth is worth when packed to its waist. A cell reading taken on a used cloth says how far into its own blinded length the cloth has gone, and so how much of the press’s pressure it is now spending, without taking the press apart for a filtration test.

The constant depends only on the cake’s packing, which is a property of the slurry, not the cloth. For one press running one duty it is fixed, and the cell reading becomes a direct record of a cloth’s condition from one wash cycle to the next.

How the readings were computed

Each cloth is a named construction in plain weave, the one weave whose holes are all one size, as only a plain weave has one size of hole found. Its hole is the sampled channel of the channel essay, which replaced the hole between four threads seen from above: four hundred and one levels through the cloth’s thickness, each with the open share left by the warp and weft beside the hole at their own heights. The monofilament reading is the integral of one over that share. The spun reading adds a yarn path at each level, the threads’ share of the slice times the yarn’s porosity, forty per cent unless stated, raised to Archie’s exponent. A used cloth has every slice from the face down to a stated depth filled with cake of forty per cent voids, conducting at that porosity to the same exponent.

Required of it, and not shown: that a monofilament reads the packed cake length to nine figures on six cloths; that a spun cloth reads short, shorter at the lower exponent, and never shorter than it is thick; that the cloth with the narrowest waist loses the most; that a used monofilament’s excess reading divided by ε−m−1\varepsilon^{-m} - 1 is the packed length at four depths on three cloths; and that a cloth packed through reads ε−m\varepsilon^{-m} times its clean reading.

What a slice-by-slice current cannot see

The current is taken straight through. Each slice is a set of parallel conductors and the slices are in series, so no current crosses sideways from a hole into a neighbouring yarn between slices. A real current spreads, which lowers the reading a little further for a spun cloth. The same assumption underlies the blinded length itself, so for a monofilament the two remain the same quantity; for a spun cloth the true reading is a little shorter still than the one drawn.

Archie’s exponent is a range, not a number. It depends on how the fibres in a yarn are arranged relative to the current, on their shape and on how the pores connect, and nothing here fixes it between 1.5 and 2. The rankings are the same at both ends of the range. The absolute shortfall is not.

The brine has to fill everything. A hydrophobic fibre, a trapped bubble in a yarn or an oily finish would close some of the yarn path without anyone intending it, and the reading would move towards LL for reasons unrelated to the cloth’s holes. A cell measurement on a spun cloth should be made after a wetting soak, and the reading should be taken as stable before it is believed.

And a used cloth’s cake has to be the cake. Fines lodged in the yarns of a spun cloth, a scale deposited by the liquor or a residue of the wash would all change the reading in the same direction as blinding. A filter is rated by the hole it does not show, and a cell rates a cloth by everything that blocks it, whether or not it came from the slurry.

Whose results these rest on

Archie’s law for the conductivity of a liquid-filled porous rock is from 1942, and the electrical formation factor has been used for the pore structure of rocks, soils and filters ever since. The analogy between Darcy flow and electrical conduction in porous media is classical, and so is the observation that the two weight pores differently. Ruth’s line and the medium resistance are from 1933 and 1935. What is put together here is the cloth: that a monofilament cloth’s brine reading is its blinded cake length exactly; that a spun cloth’s is short by an amount set by its waist and its yarn’s packing; that the cell ranks six cloths as their blinded duty does while a permeameter does not; and that a used cloth’s excess reading is its blinded length times one constant of the cake. Whether this measurement has been made on filter cloths is not known here, and nothing above has been checked against a cell.

Still open: whether a twill’s holes blind together

Every cloth above is a plain weave, because a plain weave has one hole and its blinded length is one number. Only a plain weave has one size of hole, and a twill or a satin has several. In a cell, holes of different lengths conduct side by side, so a monofilament twill reads the area-weighted harmonic mean of its hole lengths. A twill packed through with cake also passes its filtrate through all its holes side by side, by the same parallel sum, so if every hole packs, the cell still reads the blinded length exactly.

Whether every hole packs is the open part. A woven filter beats its own rating because the first particles bridge a hole at its face before they enter it, and a twill’s narrow holes would bridge at a different depth from its wide ones. The blinded duty would then be a mixture of packed depths, one per class of hole, which a clean cell cannot see.

The prediction is that a used twill’s cell reading rises in steps as its hole classes blind one after another, the narrow holes first, where a plain weave’s rises smoothly. Computing the steps needs the twill’s hole census, which exists, and a bridging depth for each class, which does not. A run of cell readings on one twill through its first few press cycles would show the steps if they are there, and a fabric to fill and a fabric to load, which found a fast and a slow path in a reinforcement, suggests they would be far apart.

Shares its objects with

Essays naming at least two of the same things, that neither author linked.

Named objects

A flat tag is an object no other essay names yet.

BlindingChannel waistFilter cakeMedium resistanceOpen areaPermeability