Detecting the clogging of a fluid filter

Abstract

A method of detecting clogging of a fluid filter is disclosed. The method includes measuring, at successive sampling instants, the pressure difference ΔP from upstream to downstream across the filter, calculating the first time derivative {ΔP}′ of the pressure difference ΔP at different instants, and then sending a clogging signal S.sub.C representative of the filter being clogged when the first derivative {ΔP}′ is equal to zero.

Claims

1. A detector device for detecting clogging of a fluid filter, the detector device being mounted in a circuit for said fluid and comprising: a sensor connected to said circuit both upstream and downstream from said filter and configured to measure, at successive sampling instants, a pressure difference ΔP across said filter; a bypass device connected to said circuit upstream and downstream from said filter and allowing the fluid to pass only when the pressure difference across said bypass device is greater than a critical value ΔP.sub.C corresponding to the filter being clogged; and a processor configured to calculate, at different instants, a first time derivative {ΔP}′ of said pressure difference ΔP, and configured to send a clogging signal S.sub.C representative of clogging of said filter when the first derivative {ΔP}′ is equal to zero.

2. A detector device according to claim 1, wherein, prior to sending said clogging signal S.sub.C, said processor is configured to send a threshold signal S.sub.S when said pressure difference ΔP exceeds a threshold value ΔP.sub.S less than the clogging pressure difference ΔP.sub.C that corresponds to said filter being clogged.

3. A detector device according to claim 1, wherein said processor is incorporated in said sensor.

4. A detector device according to claim 1, wherein said processor is configured to calculate, at different instants, a second time derivative {ΔP}″ of said pressure difference ΔP, and is configured to send said clogging signal S.sub.C only when both said second derivative {ΔP}″ and said first derivative {ΔP}′ are equal to zero.

5. An assembly comprising: a fluid circuit; and a detector device for detecting clogging of a fluid filter, the detector device being mounted in said fluid circuit, the detector device comprising: a sensor connected to said circuit both upstream and downstream from said filter and configured to measure, at successive sampling instants, a pressure difference ΔP across said filter; a bypass device connected to said fluid circuit upstream and downstream from said filter and allowing the fluid to pass only when the pressure difference across said bypass device is greater than a critical value ΔP.sub.C corresponding to the filter being clogged; and a processor configured to calculate, at different instants, a first time derivative {ΔP}′ of said pressure difference ΔP, and configured to send a clogging signal S.sub.C representative of clogging of said filter when the first derivative {ΔP}′ is equal to zero.

6. A turbomachine comprising: a fluid circuit provided with a filter; and a detector device for detecting clogging of said filter, the detector device being mounted in said fluid circuit, the detector device comprising: a sensor connected to said circuit both upstream and downstream from said filter and configured to measure, at successive sampling instants, a pressure difference ΔP across said filter; a bypass device connected to said fluid circuit upstream and downstream from said filter and allowing the fluid to pass only when the pressure difference across said bypass device is greater than a critical value ΔP.sub.C corresponding to the filter being clogged; and a processor configured to calculate, at different instants, a first time derivative {ΔP}′ of said pressure difference ΔP, and configured to send a clogging signal S.sub.C representative of clogging of said filter when the first derivative {ΔP}′ is equal to zero.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The invention can be better understood and its advantages appear more clearly on reading the following detailed description of an embodiment given by way of non-limiting example. The description refers to the accompanying drawings, in which:

(2) FIG. 1 is a diagram of a prior art filter and method for detecting clogging of the filter;

(3) FIG. 2 is a diagram of a filter of the invention and of the method of the invention for detecting clogging of the filter; and

(4) FIG. 3 is a graph showing the variation over time in the pressure difference ΔP across the filter, the variation in the first derivative of said pressure difference ΔP, and the variation in the second derivative of said pressure difference ΔP.

MORE DETAILED DESCRIPTION

(5) FIG. 2 shows a filter 20 mounted in a main branch 10 of a fluid circuit. By way of example, the filter 20 may be mounted in a fluid circuit of a turbomachine. The fluid may be constituted, for example, by a lubricant such as oil, or by a fuel. Upstream from the filter 20, the main branch 10 splits into two branches, a first branch 12 that extends the main branch 10 and that has the filter 20 mounted therein, and a second branch 14 that has a bypass device 40 mounted therein. Downstream from the bypass device 40, the second branch 14 rejoins the first branch 12 downstream from the filter 20 so as to reform the main branch 10, so that the bypass device 40 is mounted in parallel with the filter 20. In FIG. 2, the fluid flows from left to right as represented by the arrow F.

(6) A sensor 160 is connected to the first branch 12 containing the filter 20, on either side of the filter 20. Thus, a third branch 16 of the fluid circuit connects the first branch 12 upstream from the filter 20 to the second branch 12 downstream from the filter 20 via the sensor 160, which is mounted in said third branch 16. The connection of the third branch 16 upstream from the filter 20 is thus located downstream from the connection of the second branch 14 that is upstream from the filter 20, and the connection of the third branch 16 downstream from the filter 20 is thus located upstream from the connection of the second branch 14 that is downstream from the filter 20. It would also be possible to connect the third branch 16 to the main branch 10 upstream from the filter 20 and upstream from the location where the second branch 14 is connected to said main branch 10, and to connect the third branch 16 to the main branch 10 downstream from the filter 20 downstream from the location where the second branch 14 is connected to said main branch 10. Whatever the configuration, the first branch 12 including the filter 20 and the third branch 16 including the sensor 160 are in parallel, and the pressure difference across the ends of the sensor 160 is substantially equal to the pressure difference ΔP (head loss) across the ends of the filter 20. The solution shown in FIG. 2 (the third branch 16 in parallel with the first branch 12, i.e. closer to the ends of the filter 20) is preferable since the pressure difference measured by the sensor 160 is then closer to the pressure difference ΔP across the ends of the filter, given that the head loss due to the length of the branches in the circuit has no influence on the measurement. The sensor 160 is also connected to a processor 165 capable of processing a signal, typically an electrical signal. The sensor 160 is a device that is known in the prior art, as is the processor 165 associated therewith. Their structures are therefore are not described in detail below.

(7) The sensor 160 is suitable for continuously measuring the pressure difference ΔP across the filter 20 in the third branch 16. Thus, the pressure difference ΔP is measured at successive time intervals, these instants at which the measurement is taken constituting sampling over time. Ideally, these intervals are substantially regular, and the instants at which the measurement is taken are sufficiently close together to reproduce reliably the variations over time in the pressure difference ΔP. This variation in the difference ΔP is shown in FIG. 3. The pressure difference ΔP increases continuously from zero up to a value ΔP.sub.C corresponding to the filter 20 being completely clogged and to the bypass device 40 opening and allowing fluid to pass therethrough in the second branch 14. Once the filter 20 has become clogged, the difference ΔP remains constant and equal to the clogging value ΔP.sub.C.

(8) The sensor 160 sends a threshold signal S.sub.S (typically an electrical signal) to the user (e.g. the pilot) when the pressure difference ΔP across the sensor 160 (i.e. across the filter 20) reaches a threshold value ΔP.sub.S, i.e. at the instant t.sub.S in FIG. 3 (or the measurement instant immediately following t.sub.S). The threshold value ΔP.sub.S is selected to be less than the clogging value ΔP.sub.C, as shown in FIG. 3 (e.g. 90% of said critical value). Thus, when the signal coming from the sensor 160 is received by the user, the user knows that the filter 20 needs to be replaced soon. This information is forwarded to the ground crew who then perform this replacement in good time during a subsequent maintenance operation.

(9) In addition, because the pressure difference ΔP is measured continuously by sampling, the processor 165 can make use of this continuous measurement for calculating, at least each sampling instant, or at a plurality of said instants, the derivative d/dt (ΔP) of the pressure difference ΔP relative to time (this derivative is the derivative of the measurement signal delivered by the sensor 160 to the processor 165, the measurement signal being proportional to the pressure difference ΔP). This first derivative is written {ΔP}′. Thus, the derivative

(10) $\frac{d}{dt}\left(\Delta P\right)$
is calculated at an arbitrary instant t.sub.1 (written

(11) ${\frac{d}{dt}\left(\Delta P\right)]}_1$
or {ΔP}′.sub.1) can be calculated, e.g. by using two measurements ΔP.sub.1 and ΔP.sub.2 of the pressure difference ΔP at the instant t.sub.1 and at the following instant t.sub.2 by applying the following formula:

(12) $\begin{array}{cc}{\frac{d}{dt}\left(\Delta P\right)]}_1=\frac{\Delta P_2-\Delta P_1}{t_2-t_1}& \left(I\right)\end{array}$

(13) The closer together the instants t.sub.1 and t.sub.2, the more accurate the calculation of the derivative. Ideally, the instants t.sub.1 and t.sub.2 are therefore selected to be two successive sampling instants. This therefore produces a continuous measurement of the derivative {ΔP}′ of the pressure difference across the filter 20.

(14) When the filter 20 is completely clogged (i.e. at instant t.sub.c in FIG. 3), the pressure difference ΔP across the filter 20 becomes constant and remains constant (prior to the filter becoming clogged, the pressure difference increases over time, so its derivative is strictly positive). The derivative {ΔP}′ of the pressure difference then becomes zero, as shown in FIG. 3, which plots variation in the derivative {ΔP}′ as a function of time. When the derivative {ΔP}′ becomes zero (i.e. at a measurement instant equal to or immediately greater than t.sub.c), the processor 165 sends a clogging signal S.sub.C to the user who is then informed that the filter 20 has become completely clogged.

(15) Thus, by using a single sensor (the sensor 160) and a processor 165, it is possible to inform the user (and thus the maintenance crew) that the filter 20 is clogged either partially or completely, without it being necessary to use a visual indicator on the bypass device 40.

(16) The processor 165 may be incorporated in the sensor 160 or it may be separate therefrom. Under such circumstances, the processor 165 may be included in the engine monitoring unit (EMU) or in the airplane maintenance unit.

(17) In order to confirm the fact that the filter is completely clogged, it is possible to configure the processor 165 so that the clogging signal S.sub.C is delivered only when the first derivative {ΔP}′ is equal to zero on at least two successive instants. The processor 165 could calculate a first derivative of zero in error, or the calculation could take place at the extremum of a fluctuation in the pressure difference ΔP (and thus give a first derivative that is zero). The probability of such an error occurring at two distinct instants is minimal, so it then becomes certain that the filter is completely clogged if the first derivative {ΔP}′ is equal to zero at two successive instants (e.g. at two successive sampling instants).

(18) It is also possible to confirm the fact that the filter is completely clogged by calculating the second derivative

(19) $\frac{d^2}{d{t}^2}\left(\Delta P\right)$
of the pressure difference ΔP. When the first derivative is zero over a time interval of non-zero duration (and not just at a single point), the second derivative is zero over that interval. Thus, by using the processor 165 to calculate both the first derivative {ΔP}′ of the pressure difference ΔP at a given instant t.sub.1 and its second derivative (written {ΔP}″) at said instant t.sub.1, it is possible to ensure that the pressure difference ΔP is constant over a time interval around the instant t.sub.1. In present circumstances, this means that the pressure difference ΔP is constant at all times greater than t.sub.1, and thus that the filter 20 is completely clogged. It can be seen in FIG. 3 that the second derivative of the pressure difference ΔP is zero at all times t greater than the instant t.sub.c when the filter becomes completely clogged. The second derivative {ΔP}″ at an arbitrary instant t.sub.1 (written

(20) ${\frac{d^2}{d{t}^2}\left(\Delta P\right)]}_1$
or {ΔP}′.sub.1) can be calculated, e.g. by using two measurements ΔP.sub.1 and ΔP.sub.2 of the pressure difference at the instant t.sub.1 and at the following instant t.sub.2 (e.g. the sampling instant following the instant t.sub.1), and by applying the following formula together with formula (I):

(21) $\begin{array}{cc}{\frac{d^2}{d{t}^2}\left(\Delta P\right)]}_1=\frac{{{\frac{d}{dt}\left(\Delta P\right)]}_2-\frac{d}{dt}\left(\Delta P\right)]}_1}{t_2-t_1}& \left(\mathrm{II}\right)\end{array}$

(22) Thus, in order to confirm that the filter is completely clogged, it is possible to configure the processor 165 to calculate the second time derivative {ΔP}″ of the pressure difference ΔP at different instants, and for the clogging signal S.sub.C to be sent to the user only when the second derivative {ΔP}″ and the first derivative {ΔP}′ are equal to zero. This situation can only happen at instants that are later than the clogging instant t.sub.c, so it is certain that the filter is completely clogged.

(23) Alternatively, given that when the first derivative is zero over a time interval of non-zero length, the derivative of order N (where N is greater than 2) is zero over said interval, and by calculating at a given instant t.sub.1, simultaneously the first derivative {ΔP}′ of the pressure difference ΔP and its N.sup.th order derivative (written {ΔP}.sup.N), it is possible to ensure that the pressure difference ΔP is constant at any instant greater than t.sub.1, and thus that the filter 20 is completely clogged. Calculation of the N.sup.th order derivative is trivial by iterating the formula (II).

(24) The invention also provides a device for detecting clogging of the filter 20, the device comprising a sensor 160 and a processor 165 operating as stated above.