The present invention relates to a method for cleaning a liquid filter device, wherein during a first phase the liquid to be filtered is fed through a filter in a first direction, and wherein during a second phase cleaning fluid is fed in order to remove material retained in the filter. The invention also relates to a device for filtering a liquid, comprising a filtering element, pumping means and conducting means for conducting the liquid to be filtered through the filtering element, as well as means for cleaning the filter surface, suitable for carrying out at least one of the method steps as mentioned in claim 1.
A method such as described above is generally known in practice. If, after some time a certain amount of material has been retained during the filtration of a liquid, and has accumulated in some way or another in front of the filter surface, for example, in the form of a cake layer that has been deposited on the filter (with particle filtration) or a polarized layer in front of the filter surface, having an increased concentration (affinity filtration) , the quantity or the quality will diminish.
In the case of a cake layer forming on the filter surface, an increasingly higher pressure will be required in order to feed the liquid through the filter cake and the filter material . At a particular moment the necessary pressure will be so high that there is a danger of the fouling becoming lodged deep in the pores of the material. This results in an irreversible fouling of the filter material. Even backwashing will then not be able to clean the filter material . This method of backwashing consists of reverse flow feeding the filtrate through the filter material, thereby removing the filter cake deposited on the filter material. This method is generally used in practice. To this end a filtering device suitable for this method is
provided with a control unit that reverses the flow of the liquid. This may be realized, for example, by operating the pumping means such as to pump in the opposite direction or by providing separate pumping means . If a concentration of polarization layer has developed at the filter surface, it will be necessary on the one hand to raise the pressure (because of the osmotic effects) in order to maintain the quantity and on the other hand, in order to guaranty the quality of the liquid being fed through the filter material, the refreshing rate over the surface will need to be increased. At some point in time these increases in pressure and refreshing rate will technically and/or economically no longer be feasible, at which moment the material retained by the filter has to be removed. This is the technique generally used in practice. To this end such a filtering device, suitable for this method, is provided with a control unit, which batchwise, semi- continuously or continuously flushes the system with the aid of a cleaning fluid. This may be effected, for example, by displacing the system contents by means of the liquid to be filtered, such that no filtration occurs through the filter material .
The drawback with such methods known from practice is, that they are carried out at predetermined times. If the liquid to be filtered has a constant composition this is usually no problem. It is, however a problem if the composition of the liquid stream fed to the filtering element changes with time. Especially if the amount of material to be filtered from the liquid varies strongly and unpredictably, the known method has the problem that too many retained components linger in front of or on the filter material if the amount of material to be filtered from the liquid is larger than average. If, however, the amount of material to be filtered from the liquid is considerably smaller than average, cleaning will already take place without there being any real necessity. This considerably reduces the filtering capacity.
It is the object of the invention to provide an improved method of the kind mentioned in the preamble. It is a particular object of the invention to provide an improved method which, taking into consideration the type of filtered material, guarantees that the filter surface is being cleaned properly. The object of the invention is especially to provide an improved method, wherein the filter surface is cleaned at the moment that a real need arises.
In order to achieve the above-mentioned objectives, the invention provides a method as mentioned in the preamble and which is characterized in that cleaning takes place by: a. feeding filtered liquid through the filter in reverse direction, b. conducting the cleaning liquid along the filter surface to be cleaned, c. conducting the cleaning liquid and a gas along the filter surface to be cleaned; and/or d. carrying out other cleaning methods, such as using ultrasonic vibrations, chemical cleaning agents, electrical cleaning methods, and mechanical cleaning methods such as using sponge balls etcetera, while during the first period process-related operation parameters, which are an indication of the amount of material to be filtered off in the liquid to be filtered, are measured on the basis of which the total duration of time of the first period is determined. Another object of the invention is to provide an improved method, in which cleaning of the filter surface takes place on the basis of measuring operation parameters applicable to the respective type of filtration. In order to achieve this objective, the invention provides a method characterized in that the operation parameters are (preferably) chosen from: in the case of particle filtration, such as microfiltration and ultrafiltration: turbidity, concentration, pressure and viscosity of the liquid to be filtered and/or of the filtered liquid, in order to determine the difference of the parameter over the filter;
in the case of affinity filtration, such as nanofiltration and reversed osmosis filtration: conductivity, density and concentration of the liquid to be filtered and optionally of the filtered liquid, in order to determine the difference over the filter.
In particular the invention provides a method, which is characterized in that a sensor measures the relevant operation parameter of the liquid to be filtered and transmits the values measured to a control unit, on the basis of which the control unit determines the total duration of the first period.
In accordance with the invention it is possible that one control unit is provided for carrying out the various steps, but it is also possible that several separate control units are provided.
The invention provides a much improved method wherein cleaning is carried out only when it is really necessary. The point in time when cleaning takes place depends only on the fouling nature of the material to be filtered off and the liquid, i.e. the amount and the type of material to be filtered off.
The device according to the invention as mentioned in the preamble, is characterized in that the same also comprises measuring means for determining one or more operational parameters of the liquid to be filtered and/or the filtered liquid, which are an indication of the amount of material to be filtered off . Such a device allows a very economical implementation of the method according to the invention. The measuring means may consist of sensors to determine the turbidity, the concentration of material to be filtered off, the pressure of the liquid to be filtered and optionally the pressure of the filtered liquid, viscosity of the liquid, the conductivity, the density or other operational parameters that influence the filtration and in particular the supply flow of filtered off material on or in front of the filter.
Preferably, the measuring means are formed by a light sensor for determining the turbidity of the liquid to be filtered.
According to a further preferred embodiment, the measuring means are formed by sensors for determining the concentration of the liquid to be filtered and possibly the filtered liquid.
In order to facilitate the convenient implementation of the method according to the invention using the above-mentioned device, the measuring means are preferably connected with the control unit, passing the measured values on to the control unit, this latter comprising arithmetic means for calculating the total time needed for the first period, on the basis of the measured values .
According to a preferred embodiment, the measuring means comprise pressure sensors for determining the pressure difference over the filtering element, which pressure sensors are connected with the control unit and wherein the control unit comprises arithmetic means for determining the increase in pressure over the filtering element, on the basis of which the same calculates the total time required for the first period.
According to a further embodiment, the measuring means comprise conductivity sensors for determining the concentration difference over the filtering element, which conductivity sensors are connected with the control unit and wherein the control unit comprises arithmetic means for determining the difference in concentration over the filtering element, based on which the total time required for the first period is calculated.
The invention will now be illustrated with reference to some examples .
EXAMPLE 1
A filtration of a liquid generally takes place in three steps. The first step comprises the filtration, wherein a liquid is pumped through a filter material, for example, a
membrane and whereby filtration takes place. Material to be filtered off, for example, solid particles are retained on the membrane. This results in the pressure drop over the membrane, a so-called "Trans Membrane Pressure (TMP) " that forms the driving force for the process, has to be increased in order to keep the flow rate of the liquid through the membrane constant . It will be obvious that there is a limit to the TMP because the pressure will eventually be so high that irreversible fouling results. When the pressure has reached a certain value, the material filtered off that is retained on the membrane, will have to be removed. Once the residual filter material has been removed, the pressure will have returned to its minimum value, which is substantially equal to the pressure drop across the membrane alone. For removing the filtered off material, the so- called filter cake, several cleaning methods exist, such as the steps a) , b) , c) and d) to be described in Example 2. By that means the filter cake that has built up on the membrane will be removed. The filter cake can then be separated by a method known in the art. Because said filter cake has been removed, the pressure drop will at the beginning of the following filtration cycle be lower than before said backwash was performed. The time required to feed the liquid to be filtered through the membrane is called the filtration time, while the time required to clean the filter is called the cleaning time.
Finally, it is possible during cleaning to add suitable and generally known chemicals to the cleaning fluid, which for a particular length of time, to wit during soaking, are allowed to act on the membrane before being flushed away.
This chemical cleaning removes those components on a membrane that cannot be removed by an ordinary cleaning. Depending on the type of material to be filtered off, it may be necessary to perform a chemical cleaning at each cleaning or only per two cleanings, three cleanings, or more cleanings, for example, once every ten cleanings.
According to the invention the time required for filtration may be based on the type and the amount of
pollution in the liquid. According to the invention, the amount of material in the liquid to be filtered off is determined for this purpose.
According to a first embodiment, the turbidity of the liquid to be filtered is determined. According to a second embodiment, the rate at which the TMP increases is determined. These two methods give a very good indication of the extent to which the liquid to be filtered fouls the membrane . By measuring the turbidity of the liquid to be filtered, it is possible to determine how frequently the filtering device has to be cleaned in order to maintain a suitable activity of the filtering device. If the turbidity of the liquid to be filtered increases, the duration of filtration, the filtration time, has to be reduced. In contrast, if there is less turbidity, the filtration time may be prolonged.
According to a further embodiment it is possible to use a minimum filtration time and a maximum filtration time. If the liquid to be filtered is only slightly turbid it is still possible to perform backwashing at predetermined points in time.
If the filtration time is determined on the basis of the increase in TMP, the amount of filtered off material in the liquid to be filtered is determined. The rate at which the TMP increases is determined by the slope of the TMP / time graph, as illustrated in Figure 1. The derivative of the tangent of this graph is in that case used as basis for the determination of the total filtration time. If this derivative is greater, the filtration time will be shortened; in contrast, if this derivative becomes smaller, the filtration time will be prolonged. Here it is again possible, as mentioned before, to in advance input a maximum and minimum filtration time to the control unit, so that the maximum filtration time is predetermined even if the derivative of the graph has a very low value.
EXAMPLE 2
In this example a liquid to be cleaned is cleaned by means of affinity filtration. In general, affinity filtration may consist of reverse osmosis or nanofiltration. A liquid to be cleaned is filtered in a semi-dead-end filter comprising a reverse osmosis or nanofiltration module. When performing semi-dead-end affinity filtration, the concentration of the substance to be filtered off at the feed side of the membrane surface will eventually become so high that the membrane's apparent retention becomes too low. The result is that the quality of the permeate diminishes. In contrast with particle filtration, there is substantially no formation of filter cake with affinity filtration, but the concentration of a substance to be filtered off will increase at the filter surface. In the case of affinity filtration therefore, cleaning the filter surface means flushing the liquid having an increased concentration of filtered material away from the filter surface. Such cleaning can be performed by a. feeding filtered liquid in reverse direction through the filter, b. feeding cleaning liquid across the filter surface to be cleaned, c. passing cleaning liquid and a gas across the filter surface to be cleaned; and/or d. performing other cleaning methods, such as using ultrasonic vibration, chemical cleaning agents, electrical cleaning methods, and mechanical cleaning methods such as the use of sponge balls etcetera. With affinity filtration, the increase in concentration of a particular measurable component is measured during filtration. Based on how the concentration increase proceeds it will be decided to carry out cleaning when a particular maximum value is reached. The parameters that are measured in affinity filtration may include, for example, conductivity, density and concentration of the liquid to be filtered. Optionally, the differences between these values in the liquid to be
filtered and the permeate may also be measured. Based on how the difference measurement proceeds, it may be determined when cleaning has to take place.
It may be preferred for the regulation to comprise in any case a minimum filtration time and possibly also a maximum filtration time. The control unit may, for example, comprise instructions to the effect that the filtration time is to take at least ten minutes. Of course, any other length of time is also possible. This will depend solely on the nature and the amount of pollution in the liquid to be filtered.
If the filtration time is determined on the basis of the increase of the conductivity (G) , the concentration of the filtered off material in the liquid to be filtered will be determined. The rate of increase of the conductivity is determined by the slope of the conductivity (G) / time graph, as shown in Figure 2. In that case the derivative of the tangent of this graph is used as basis for the determination of the filtration time. If this derivative increases, the filtration time will be shortened; In contrast, if this derivative decreases, the filtration time will be prolonged. Here also it is possible to previously enter a maximum and minimum filtration time into the control unit in order to, as mentioned above, establish the maximum filtration time even at a derivative of very low value.
In accordance with the invention a much improved method for cleaning a liquid filter is provided. By this means the problem of a liquid filter being cleaned too frequently is avoided. According to the invention, an increase of available filtering time is obtained. This applies in particular when there is only a small amount of material to be filtered in the liquid.
If, in contrast, there is much material to be filtered in the liquid, the method according to the invention prevents the development of extreme processing conditions such as too high a TMP or too quick a refreshing rate, which could result in irreversible fouling or damage to the filter.
Therefore, the advantage of the method according to the invention is to be found in the fact that the filtration time completely and only depends on the nature and the amount of the material to be filtered off in a liquid to be filtered.