NZ625041B2 - A method for accelerated testing of a membrane module - Google Patents

Abstract

Disclosed is a method for accelerated testing of a membrane module for resistance to cyclic stresses during operation of an industrial membrane separation process comprising loading at least one membrane module into a test cell of a test plant; conducting said industrial membrane separation process while subjecting said membrane module to cyclic stresses during a plurality of membrane operating cycles, each cycle being of significantly lesser duration than used for operation of said industrial membrane separation process in an operating plant; and testing said membrane module for component failure caused by said cyclic stresses. while subjecting said membrane module to cyclic stresses during a plurality of membrane operating cycles, each cycle being of significantly lesser duration than used for operation of said industrial membrane separation process in an operating plant; and testing said membrane module for component failure caused by said cyclic stresses.

Description

A METHOD FOR ACCELERATED TESTING OF A MEMBRANE MODULE This invention relates to a method of accelerated g of a membrane module.

Membrane processes are commonly used in industry for separating components from fluid mixtures, whether the fluid is in liquid or gaseous form.

Membranes of semi-permeable material enable separation of ents as follows. Some components of the fluid mixture permeate through the ne, though at different rates. Other ents will not pass through the membrane at all. Such components are said to be retained by the ne.

In the field of water ent, for example, membrane processes may be used to purify an “effluent” water stream containing water in admixture with contaminant components. Such industrial processes involve membrane tion of the contaminant components from the effluent water, allowing the treated water to be re-used by industry or discharged safely into the environment.

An example of such a process, advantageously involving a combination of membrane separation processes (nanofiltration and e osmosis) is disclosed in the Applicant’s WO 2010037156 “Process and Plant for Treating a Water Stream”, the contents of which are hereby incorporated herein by nce.

Membrane processes may be used to achieve separation of substances in many applications and process media including water treatment for desalination, filtration and in membrane bioreactors, which are a specialised type of filtration application.

Industrial application of membrane processes such as disclosed in WO 2010037156, requires suitable equipment in the form of membrane modules for which s designs are ble: hollow fibre, tubular, spiral woundgand — less commonly plate and frame. Hollow fibre membrane modules are often red because hollow fibre membranes allow the highest membrane surface area per unit volume of a vessel to accommodate them. A hollow fibre membrane module typically comprises a bundle of the hollow membrane fibres sometimes odated within a pressurisable housing or immersed in an open tank configuration. A hollow fibre bundle extends between two opposed headers or "pots" forming part of the module housing and the opposite ends of each hollow fibre are potted and sealed within each “pot”, typically of polymer or resin al (particularly epoxy or polyurethane resin), to form a tube sheet.

The ing design is analogous to a shell and tube heat exchanger. The fluid to be treated is contacted with the hollow fibres such that components are separated, with liquid lly passing inwards to the hollow centre of the fibre under the action of a moderate pressure gradient, and the solids ing outside the fibre.

Membrane separation ses often e a number of such membrane modules operating in parallel. A number of such modules may comprise a rack which is located, either in a pressurised system or submerged in a tank style water treatment cell to which water for treatment has been introduced to remove contaminants from the water, for example in a membrane bioreactor.

The water treatment cell is lly agitated by sparging air or other gas through it and or by recirculating a liquid flow through it. Such agitation and or recirculation prevents solids buildup on ne fibres that would impair membrane treatment performance.

In an industrial plant, such as a waste water treatment plant (but not by any means limited to such plant), the selected membrane modules are likely to 2O have a significant capital cost. At the same time, membrane service life of several years is likely to be required. A constraint on membrane service life is membrane module failure due to repeated cyclic stresses imposed on the hollow fibre membranes and supporting components such as the pots during operation.

In normal operation, hollow fibre membranes are ted to filtration, relaxation and pressure testing . Treated fluid must be extracted from the hollow fibre membranes by means of a pressure differential. This can be by means of suction in the case of submerged fibres, or by pressure difference in pressurised systems. In each case, a pump or source of pressure difference is "on" during a filtration cycle, which applies substantial bending forces and cyclic stresses on 3O the membrane module system, particularly at the pots in which opposed ends of the hollow fibre membranes are sealed. The pots are also likely to be subjected to other cyclic ical stresses such as periodic stresses in the reverse or pressurised direction during a pressurisation cycle. Periods when the pump or other source of pressure means is "off" are called relaxation cycles. Such relaxation cycles may have the same on as filtration cycles or may be much r. In water treatment plants, sludge removal operations, or backwashing, may be conducted during relaxation cycles, for example through agitating mixed liquor surrounding the hollow fibre membranes, shaking them to remove .

In addition, pressure integrity testing/pressure decay cycles may be conducted in which the hollow fibre membranes are pressurised with compressed air and dynamic behaviour is observed in order to assess membrane integrity and absence of leaks. Decay or decline in pressure (for example) over a pre-set time provides a direct measure of ity. Other pressure decay tests can use a vacuum source and observe rate of increase of pressure. In either case, if the decay gradient is too steep, particularly being steeper than a pre—determined gradient consistent with membrane integrity, the membrane has unsatisfactory integrity and is not fit for purpose.

Daily operation of a membrane separation process within a water treatment plant may lly be designed to e, for e and without any intent of limitation, over one hundred filtration cycles (about 14 minute design cycle time) and one pressurisation cycle per day. 2O Over a period of time, the continuous imposition of cyclic stresses — possibly in combination with corrosion/degradation caused by chemical agents for example used for disinfection (for example sodium hypochlorite) — can lead to damage to the membrane , particularly through cracking and fatigue failure. If cracking, particularly at the pots or tubesheets, occurs the membrane module is typically no longer serviceable. This is a very significant failure mode, particularly in ne bioreactors, which can occur despite the membrane material for the hollow fibres being ise correctly selected from the perspective of mechanical properties (see, for example, ess, AE et al, Mechanical analysis of hollow fibre membrane integrity in water reuse 3O ations, Desalination 180 (2005), 5-14). The reason is that separation of fluid to be treated from treated fluid: for example mixed liquor (dirty water) from clean water in the context of water treatment is no longer possible. The ne module must be discarded.

As alluded to above, it may be several years before a membrane module becomes unserviceable. However, the exact timing of failure (service life) is uncertain and this makes planning and budgeting for a water or wastewater treatment plant difficult. Such uncertainty also impacts selection of a membrane module which is best suited for the ne separation process. All other factors being equal, the plant designer will select a membrane module having the longest service life. If it takes several years, for example 5 years or more, to determine likely failure time, through statistical is, it is very ult for an optimal selection to be made. Whilst manufacturer and brand provides a guide to membrane module service life, greater certainty is required. The membrane module manufacturer is not in an ideal position to provide an accurate service life prediction (and likely will not e one). The reason is simple in many cases.

The specific operating nment for a particular industrial ne tion process often cannot be reasonably replicated by a ne module manufacturer ing its usual testing protocols. A real challenge for the membrane module manufacturer is duplicating the often time variant compositional nature, and - to a perhaps lesser extent - properties (such as temperature) of fluids subjected to membrane separation. For example, effluent subjected to membrane separation in a water treatment plant often has a highly dynamic, that is time and/or seasonally variant, composition and flow rate.

Properties such as temperature may also vary with time and the nature of the effluent being treated by the water treatment plant. ne module service life predictions based on tests using sample, or synthetic, effluents of near-static or near-constant composition is simply not sufficiently te for budgetting purposes.

It is an object of the present invention to provide r accuracy of prediction of membrane module service life than currently available. 3O With this object in view, the present invention provides a method for accelerated testing of a membrane module for resistance to cyclic stresses during operation of an industrial membrane separation process comprising loading at least one membrane module into a test cell of a test plant; conducting said ne separation process while subjecting said membrane module to cyclic stresses during a plurality of membrane operating cycles, each cycle being of significantly lesser duration than used for operation of said industrial membrane tion process in an operating plant; and g said ne module for ent failure caused by said cyclic stresses.

Such testing, which may usefully include imposing chemical treatments similar to those used in the industrial membrane separation process (though including testing under more aggressive chemical conditions), allows a mean service life for the membrane module to be predicted, advantageously in a much lesser timeframe than actual service life. Such information may be advantageously employed to assist in membrane module selection and cost budgeting for a membrane separation plant.

The method accelerates ne module testing because the test cycle duration is significantly shorter than design duration, the duration to be used for membrane operating cycles in the operating plant. Consequently, frequency of membrane operating cycles applied during testing is significantly greater than the design frequency, that is, the ncy at which stress cycles will be applied to the membrane module in normal industrial service. industrial service is not limited to commercial e. Practice of membrane tion processes as part of a public utility is also included within the scope of industrial service.

Advantageously, the test plant forms part of, or is connected to, an actual operating plant implementing the industrial membrane separation process. This integration of test plant and operating plant enables testing of ne s under actual industrial ing conditions, rather than simulated or synthetic conditions in an artificial environment, addressing inaccuracies created by failure to test membrane modules under dynamic compositional, property and flow rate ions typically encountered in industrial plants. The results of testing are substantially more reliable than tests performed distant from an operating plant implementing said industrial membrane separation process, potentially on bench or very small scale, with fluid s for membrane sing of near-constant or static quality and rather than time variant quality.

Membrane operating cycles, which are typical and of particular though not limiting importance, may comprise filtration cycles; relaxation cycles; and/or pressure decay testing cycles in which the membranes of the membrane modules are tested for leaks. That is, the membrane modules may only be subjected to tion and relaxation cycles; or pressure decay testing cycles. Alternatively, membrane modules may be subjected to both filtration and tion cycles and pressure decay testing cycles. Filtration cycles correspond with s when pump means forming part of the test cell are operated, or turned on, to apply pressure (positive or negative) to a membrane. Relaxation cycles correspond with periods when pump means forming part of the test cell are turned off.

Membrane cleaning operations, such as backwashing or sludge removal operations, may be conducted during relaxation cycles. These cycles are significantly rated over operating plant cycles for test purposes.

The test plant may include a plurality of test cells. in such cases, test cells i5 may be operated in parallel with one test cell imposing different membrane operating cycles and a different cyclic stress regime on membrane modules than in another test cell. Chemical testing, to assess corrosion/degradation phenomena, may also be practised in one or more test cells. Such parallel test cell operation accelerates the collection of test data.

The frequency of filtration and relaxation cycles to which a membrane module is subjected during testing is accelerated to a frequency which would lly be icantly greater than the frequency of any pressure decay/pressure integrity g cycles used in the operating plant, for example up to and greater than 100 filtration cycles to 1 pressurisation cycle. This can induce significant pressure shocks at pumps and valves which operating plants would normally take steps to prevent (see, for e, Huisman, IH and Williams, K, Autopsy and failure analysis of ultrafiltration membranes from a waste-water treatment system, nation 165 (2004), 161-164)). The g method, in st, is intended to purposefully induce such pressure shocks to increase testing severity.

The membrane module to be tested conveniently comprises hollow fibre membranes as such membranes are often selected for industrial membrane separation processes for reasons described above. In this case, the ne module comprises a housing accommodating a bundle of the hollow membrane fibres, each hollow fibre ing between two opposed headers (also known as pots) with the opposite ends of each hollow fibre being potted and sealed within a header, typically of polymeric or resin, such as epoxy or polyurethane resin, material. These pots or headers are critical components of the membrane module. Component failure would be indicated; in particular, when cracking of a pot or adhesion failure n the fibres normally retained in the pot, and the pot, occurred during testing. The membrane module testing method may also be applied to other membrane module types such as tubular, spiral wound and plate and frame membrane module types.

During testing, a plurality or rack of membrane s may be loaded into the test cell to enable simultaneous testing. The resulting increase in data accelerates prediction of the likely time to failure (i.e. service life) of a membrane module in service. The prediction of likely time to failure should not require any complicated atical modelling.

The method may comprise subjecting each membrane module to a testing cycle having a predetermined number of stress cycles to be imposed over a predetermined period of time. The predetermined number of stress cycles may correspond with required minimum service life (typically measured in years) though this number of stress cycles will be imposed in a much shorter period (in the order of weeks or months).

The membrane module testing method may be used for other purposes.

During testing, duration of a filtration cycle will be icantly r than during normal service. Such extremely short filtration cycles may increase filtration flux over normal service. Filtration flux during g may thus be ed and ed with filtration flux during normal e.

Alternatively, or additionally, testing may be used to assess whether resistance to membrane fouling is improved by r filtration cycle duration.

The membrane module testing method may be used to test membrane modules and predict membrane module service life for a range of different membrane separation processes. However, the testing method is particularly ageous for industrial ses in which typical required membrane service life has a duration of a plurality of years.

Liquid treatment processes, especially water treatment processes, are a ularly important example of this ry of processes since membrane separation processes are widely used for separation of inants from water.

In addition, membrane modules for use in large volume water treatment plants have a typically high capital cost. The testing method is especially suitable for testing membrane modules to be used in membrane bioreactors, especially under submerged conditions. For testing this application, test cell(s) are configured to simulate membrane bioreactor(s) and membrane modules, typically comprising hollow fibre membranes are submerged in mixed liquor in the test ).

The testing method may be applied to any waste water treatment processes ing those where a membrane module will be subjected to a temperature greater than about 2590, or even greater than 3090 and approaching 40%. At such temperatures, which may ularly be encountered with effluent water s containing effluent from an industrial process, membrane module operating temperature may be greater than the glass transition temperature of the polymer or resin material of the headers of the membrane modules Thus the g method may be applied at ent temperatures to assess dependence of service life on service ing temperature.

Testing severity may also be increased by imposing various chemical ents, such as sodium hypochlorite solutions typically used as disinfectants in water treatment processes, at a plurality of solution strengths, for varying durations so as to simulate the type of chemical attack and associated degradation that might occur during normal use over a period of years. Such chemical treatments could also be imposed where the membrane operating cycle duration is the same as, or similar to, duration used in the operating plant.

An embodiment of the method of accelerated testing of a membrane module in accordance with the present invention will now be described with reference to the accompanying figures in which: Fig. 1 is a tic process flowsheet for a water ent plant ing a membrane module of type to be tested in accordance with the method of testing of the invention.

Fig. 2 is a schematic illustration of a membrane module used in an ultrafiltration stage of the water ent plant of Fig. 1.

Fig. 3 is a s flow m for a test plant for testing membrane modules in accordance with the method of the present invention.

Fig. 4 is a schematic process flowsheet for a water treatment plant ing a test plant for which the process flow diagram is provided as Fig. 3.

Referring now to Fig. 1, there is shown a water treatment plant 100 comprising two bic reactor stages (AFH and AR2) to produce treated water and biogas, a portion of which may be combusted in a gas turbine to generate power for the plant. Water treatment plant 100 treats up to 35 million litres of waste water per day. ARt and AR2 supply pre-treated influent water to aeration basins and membrane bioreactor (“MBR”) systems (AB1/MOSf and ABZ/MOS2).

The anaerobic reactors treat raw influent water which is a combination of domestic nt and industrial effluent received from a Kraft process pulp and paper mill through a dedicated waste line through a digestion process. This effluent or waste water has contaminant ents as measured by a high biochemical oxygen demand (”BOD”) and dissolved organic carbon (“DOC”) content which provides a strong brown colouration to the waste water.

The treated effluent from the anaerobic reactor stages AW and AR2 passes to the aeration basins A81 and A82 respectively. These basins provide aerobic digestion conditions to further treat the waste water. The mixed liquor (“dirty water”) is recirculated from aeration basins A81 and A82 into ultrafiltration (“UF”) membrane operating systems M081 and M082 each of which includes a large number of membrane modules 40 as schematically illustrated in Fig. 2.

Water treatment plant 100 could comprise a large number of such UF membrane operating systems (M08) and several thousand membrane modules 40 for UF separation of the ent represented by high BOD. The resulting ultrafiltrate or permeate has nearly undetectable levels of BOD. Each combined AB+MOS system comprises a membrane bioreactor (“MBR”) and the ultrafiltrate may be ed to as MBR filtrate or permeate.

MBR permeate is directed to nanofiltration process unit 20 for removing DOC from the MBR permeate to e a water quality acceptable for feed to reverse s (“RO”) process unit 30. The RO process unit 30 performs desalination (especially removing te) and allows production of a water stream that may be ed to the pulp and paper mill with benefit for the process economics of that mill.

Fig. 2 shows a schematic, referenced at 1O http://www.separationprocesses.com/Membrane/MT_FigGen15.htm, of an ultrafiltration (“UF”) ne module 40 suitable for use in the M081 and M082 units. UF membrane module 40 is of hollow fibre type comprising a bundle 44 of hollow fibre membranes 41 accommodated within a pressurisable housing 46. Bundle 44 s between two vertically opposed headers or pots 42 and 43 forming part of the module housing 46. Headers 42 and 43 are of epoxy resin. The opposite ends of each hollow fibre of bundle 44 are embedded and sealed within headers 42 and 43 respectively. The resulting design of membrane module 40 is analogous to a shell and tube heat exchanger with feed mixed liquor 47 entering the housing 46 (shell side) ting the hollow fibre membranes 41 with “clean” water, essentially free of BOD, permeating the membranes 41 to enter the lumens 45 of membrane fibres to be recovered as a permeate stream 49. Water relatively concentrated in BOD is recovered as a retentate stream 48. The flow 51 of permeate stream 49 is counter-current to the flow of the feed mixed liquor 47 through the housing 46.

During typical daily (24 hour duration) operation of membrane modules 40, the bundle 44 of hollow fibre membranes 41 is subjected to about 110 filtration cycles in which permeate 49 is recovered, whilst corresponding pump means are operated, by suction from the hollow fibre membranes 41. Therefore, each filtration cycle, which imposes the cyclic stress above described, has a design 3O duration of approximately 14 minutes. Following a filtration cycle, pump means is turned off during a relaxation cycle (also of 14 minute design on). Sludge removal ions are conducted throughout this 14 minute cycle time, but particularly facilitated by the relaxation step, for example by agitating mixed liquor surrounding hollow fibre membranes 41 to shake the membranes and remove adherent sludge. The sludge so d is typically carried away by a recirculation flow.

In a distinct testing cycle, med at lesser ncy than filtration and relaxation cycles, the bundle 44 of hollow fibre membranes 41 is pressurised by compressed air in a cycle called a pressure integrity testing (“PIT”) cycle or pressure decay g (“PDT”) cycle which also imposes a cyclic stress on the membrane modules 40. PIT or PDT testing is used to check whether hollow fibre membranes 41 leak. In a PIT/PDT test, hollow fibre nes 41 are pressurised to a predetermined pressure (for example 105 kPa) with compressed air. Decay of the hollow fibre membrane 41 pressure with time indicates whether leaks exist in the hollow fibre nes 41. Such leaks are issible since they compromise the extremely high mechanical disinfection, and reduction in biota ially pathogenic bacteria and protozoa) which is required during a water treatment s.

The filtration and PIT/PDT cycles impose cyclic mechanical stresses on the membrane module 40. In particular, the headers 42 and 43 of epoxy or polyurethane resin are subjected to bending forces which will cause failure through stress cracking. When such cracking occurs, mixed liquor will mix with permeate and the membrane module 40 is no longer serviceable. Membrane module 40 must then be replaced. Whilst it is ted that such failure will occur at a mean service life of about 5 years (approximately 200000 filtration cycles), there is considerable uncertainty over this and cost/capex budgeting for replacement of membrane modules is y uncertain.

Fig. 3 shows schematic of a test plant 60, which — as shown in Fig. 4 — forms part of an operating water treatment plant 100 located in and, Victoria, Australia. Water treatment plant 100 has capacity to treat a very high volume of aqueous effluent comprising domestic and industrial effluent (particularly pulp effluent from a Kraft paper mill) having different and time variant compositions, colour and temperature. For example, the domestic effluent contains relatively low levels of sulphate whereas the industrial effluent contains relatively high levels of sulphate and typically has a higher temperature.

The domestic and industrial effluents are mixed into a mixed liquor for treatment within the water treatment plant 100. Having test plant 60 forming part of operating treatment plant 100 enables testing of membrane s 40 under rial operating conditions. So, the results of testing are more reliable than tests performed distant from the water treatment plant 100, potentially on bench or very small scale, with effluents of constant rather than time variant quality.

The test plant 60 is lled by the water treatment plant 100 control system (PLC/SCADA control system).

Test plant 60 is used for implementing e g protocols in which ne modules 40 are subjected to cyclic stresses during a plurality of ne ing cycles, each cycle being of significantly lesser duration than used for ultrafiltration using the MOS units; and testing the membrane s 40 for component failure caused by said cyclic stresses. In particular, the stress cycles are imposed on the membrane modules 40 by the filtration (suction), relaxation and pressurisation cycles described above. Such g allows a mean service life for the membrane module 40 to be predicted by acceleration mechanical stress testing in a period significantly less than estimated service life.

Such information may be advantageously employed to assist in membrane module selection and cost budgeting for water treatment plant 100 and, in particular, the UF membrane operating systems (MOS units exemplified by M081 and M082).

Test plant 60 includes two test cells 61 and 70, each configured as a membrane bioreactor in which ne modules 40 are submerged while conducting the ne separation process. Each test cell 61, 70 is loaded with a rack of 16 membrane modules 40 for parallel testing, such testing involving ultrafiltration of mixed liquor for removal of BOD as described above.

Simultaneous testing of a icant number of membrane modules 40 allows 3O data acquisition to be accelerated further as sample size is increased and mean membrane module 40 service life can be predicted with greater accuracy. It is to be understood that membrane modules 40 are not expected to fail at the same time. Rather, different membrane modules 40 will have different endurance to stress cycles and fatigue failure and the output of testing is typically a statistical distribution of duration to failure, this tical bution having a mean duration to failure.

Test cell 61 comprises three filtrate pumps 62 operable in parallel dependent on the vacuum pressure that is required for the test cell 61. One or more filtrate pumps 62 are therefore used dependent on the required operating pressure. Higher suction pressures will require more of the pumping capacity to be used. Filtrate pumps 62 operate to draw UF permeate, at a temperature about 3090, from a single rack of 16 membrane modules 40 located in test cell 61 during a filtration cycle. Filtrate is sent to test filtrate storage tank 64 A submersible pump 66 in the filtrate storage tank 64 returns filtrate back to the test cell 61 based upon liquid level in the filtrate storage tank 64. The required filtrate pumps 62 e continuously. Filtrate flow is controlled by two actuated valves 68 and 69, conveniently of solenoid type.

Test cell 61 is also configured to enable pressure decay/pressure ity PIT”) testing, that is testing of the bundle 44 of hollow fibres for impermissible leaks as above described. To that end, air compressor 63 supplies compressed air through line 162 and automated control valve 163 to pressurise membrane modules 40 when required during a T test. Alternatively, compressed air may be supplied from the compressed air supply for water treatment plant 10.

Test cell 70 operates identically to test cell 61 except that it is configured, in accordance with membrane module 40 test strategy, not to perform PDT/PIT testing. Test cell 70 therefore subjects its rack of 16 ne s 40 to filtration and relaxation cycles only. Test cell 70 has three variable speed filtrate pumps 72 (of progressive cavity type) le in el operable in parallel dependent on the vacuum pressure that is ed for the test cell 70. One or more filtrate pumps 72 are therefore used, dependent on the required operating pressure, to draw filtrate from a single rack of 16 membrane modules 40 located in test cell 70 during a filtration cycle. Filtrate is sent to filtrate e tank 74.

Submersible pump 66 in the filtrate storage tank 64 s filtrate back to the test cell 70 based upon liquid level in the te storage tank 64. The required filtrate pumps 72 operate uously with filtrate flow being controlled by two actuated valves 78 and 79, conveniently of solenoid type.

Expected filtrate flow range from each rack is in the range 7.5 to 25 m3/h and operating trans membrane pressure (TMP) during a filtration mode is 45 kPa.

Filtrate flow range to test cell 60 or 70 via submersible pump 64 may be up to 50 m3/h.

Test plant 60 may be operated, by way of example, as follows. Filtration test cycle on for both test cells 60 may be pre-set at 10 seconds, this duration being very significantly less than the 14 minutes period required in normal service of water treatment plant 10. The filtration cycle duration is the duration for which filtrate pumps 62 and 74 draw filtrate from the rack of membrane modules 40. Filtrate pumps 62 and 74 are controlled to the t duration for the filtration cycle. A tion cycle is followed by a relaxation test cycle also having a pre-set duration. In this case, duration of the filtration and relaxation cycles are the same and set at 10 seconds for g. This test cycle duration is purely by way of example but it is significantly less than the about 14 minute duration of filtration and tion cycles during normal service of M081 and M082 of the water treatment plant 100. Consequently, the frequency of stress cycles imposed on the membrane modules 40 is accelerated to ncy being very significantly greater than during normal service of M081 and M082 of water treatment plant 100.

During the filtration cycle (10 seconds duration for one test protocol), actuation valve 68 is open and actuation valve 69 is closed. The three filtrate pumps 62 are here operated to achieve a vacuum re of -45 kPa, measured and controlled by pressure transmitter 67. Flow is also measured by flow meter 65 but will not be the lling parameter.

At the completion of the filtration cycle, actuation valve 69 will open and mixed liquor will be drawn from the filtration storage tank 64 during a relaxation 3O cycle. 0.5 seconds later, actuation valve 68 will close. During the relaxation cycle (10 seconds duration for one test protocol), filtrate pumps 62 operate at the same speed as during the filtration cycle (that is pump speed remains on hold). A different speed control program for filtrate pumps 62 could be adopted.

At the completion of the relaxation cycle, actuation valve 68 will open and 0.5 seconds later, actuation valve 69 will close. The intention is that at no time are both actuation valves 68 and 69 closed and that when swapping from filtration to relaxation and back that for a short period of time both actuation valves 68 and 69 are open. This is required as filtrate pumps 62 are progressive cavity type.

Further, the testing protocol is very aggressive by its nature and such short cycle times put significant pressure on valves and pumps as well as, potentially, on the 1O membrane modules 40. Pressure shocks are quite likely in contrast with operating plant. These pressure shocks can be ed intentionally to se the test severity, or can be minimised to reduce it down to a level more resembling normal operational pressure exposure. The test filtration and relaxation cycle times are ore ideally to be selected to produce failure data that is relevant to ting membrane module 40 service life and rather than failure data that s simply from the testing regime.

Test cell 61 is also configured for PDT/PIT testing. At a time determined by an operator of the test plant 60, te pumps 62 are turned off and a PDT is performed by opening control valve 163 and ing compressed air through 2O line 162 for delivery to membrane modules 40. The PDT/PIT test mimics the test as required, by regulation, for M081 and M082 of water treatment plant 100.

Compressed air is delivered to the membrane modules 40 for a pre—set pressurisation duration (60 seconds for one testing protocol) in order to pressurise the hollow fibre membranes 41. At completion of the pressurisation duration (60 seconds), provided that a ient start pressure is obtained, the PDT/PIT test begins. Control valve 163 is closed and drop in re of the hollow fibre membranes 41 is measured over 60 seconds. This on is also for example only. Different durations could be selected for g. The water treatment plant 10 control system (SCADA network) records the results. At 3O conclusion of the PDT/PIT test, the test cell 61 may once again be returned to testing during filtration and relaxation cycles. However, test cell 61 could also be used simply to perform PDT/PIT tests to determine the influence of such testing on membrane module 40 e failure. In this way, test plant 60 can test for the most likely cause of membrane module 40 failure, cyclic stresses due to filtration/relaxation cycles or cyclic stresses due to PDT/PIT testing or a combination of the two types of cyclic stress. lf PDT/PIT testing is a cause of failure, this has ations for use of such testing in water treatment plant 100.

Test cell 61 is returned to filtration/relaxation testing following removal of air (typically though, due to the possibility of hollow fibre membrane 41 failure, not always) from the hollow fibre membranes 41 of membrane modules 40 (see Fig. 2). in order to do this, solenoid valve 165 is opened and service water passed 1O through the venturi 164 to provide motive force. This differs from the protocol used in water treatment plant 100 where a compressed air supply is used to provide motive force to venturi(s) in M081 and M082. After a preset time, solenoid valve 166 is opened and a vacuum is d to the hollow fibre membranes 41. This vacuum removes excess air and draws an quid mixture into line 167. This air-liquid mixture is then discharged into test cell 61. After a t duration (60 seconds for example), solenoid valves 165 and 166 are closed and test cell 61 is ready for restart of tion/relaxation cycle stress testing.

Test cell 70 involves similar operation to test cell 61 except that PDT/PIT 2O testing is not performed. Test cell 70 may be used to test membrane modules 40 with different filtration/relaxation cycle duration (say 20 s test cycle duration rather than 10 seconds) being used during testing. The actuating valves 78 and 79 are operated in the same manner as actuating valves 68 and 69. re transmitter 77 takes the place of pressure transmitter 67 and flow meter 75 takes the place of flow meter 65.

An indicator of hollow fibre membrane 41 failure during filtration/relaxation cycle testing is increased turbidity downstream of test cells 61 and 70. ity is therefore monitored as an indicator of possible failure. Membrane modules 40 can then be inspected to assess r failure has occurred and location of any cracks in the membrane module headers ("pots") 42 and 43.

Tests may involve testing with waste waters at different temperatures.

Though failure is less likely with domestic effluent water having temperature of -2596, membrane module 40 g in that temperature range may be conducted. Tests at higher temperature, and in particular in the temperature range 30-3990 typical of conditions where the waste water contains waste water from the Kraft process pulp mill are also performed. Membrane operating cycle duration and frequency may be varied during testing dependent on waste water temperature.

Testing severity in test cells 61, 70 may also be increased by imposing various chemical treatments, for example sodium hypochlorite solutions typically used as disinfectants in water treatment processes, at a plurality of solution strengths, for varying durations so as to simulate the type of chemical attack and associated degradation that might occur during normal use over a period of years. Such chemical treatments could also be imposed in test cycles where the membrane operating cycle duration is the same as, or similar to, duration used in the ing plant.

Different membrane modules 40 will endure stress testing for ing s. At the end of testing, a tical distribution of number of fatigue failures of pots 42, 43 to time to failure can be ed and a mean time to failure (equated to predicted service life e the number of test cycles to failure in testing plant (duration of 20 seconds = 10 seconds tion/10 seconds relaxation) can be compared with the about 200000 cycles equating with 5 years service (or 40000 cycles per year)) of a membrane module 40 predicted using conventional statistical techniques. A table showing conversion from testing duration to predicted service life is as follows: Table 1 Conversion — Accelerated Testing Duration to Service Life Number of Cycles g Duration (weeks) Predicted Service Life (weeks) (years) 50000 1.6 1.25 75000 2.4 1.88 100000 3.3 2.5 125000 41 3.1 150000 50 3.75 200000 6.6 5.0 As can be seen, mean service life can therefore be predicted in highly accelerated manner in 6 to 7 weeks using test plant 60 (based on 100% plant availability) rather than the 5 plus years it could take if membrane modules 40 are tested in normal service of water treatment plant 10. Such accelerated testing enables more confidence in membrane module 40 selection and ing for water treatment plant 100, particularly in terms of membrane module 40 replacement costs.

Test plant 60 may be used for other purposes. Above was described an extremely short filtration/relaxation cycle duration (10 seconds each for filtration and relaxation). Such extremely short cycle duration may impact either positively or negatively on membrane flux and membrane tion efficiency. Test plant 60 allows the effect of filtration/relaxation cycle duration on separation efficiency to be investigated.

Alternatively, or additionally, testing plant 60 may be used to assess whether resistance to membrane fouling is improved by r filtration cycle duration. Fouling is a icant problem in many membrane separation processes, particularly those used in waste water treatment, and most notably within membrane bioreactors. If filtration cycle duration can be used to improve resistance to fouling, and perhaps result in lowered usage of cleaning chemicals or other fouling ents such as agitation and / or recirculation flow, this would also have benefits for water treatment plant to economics, either by permitting reduced initial capital investment, or ng increased hput capacity with vely minor equipment modifications (fast acting valves, and more precise controls etc). cations and variations to the method for accelerated testing of a membrane module described in this specification may be apparent to the skilled reader of this sure. Such modifications and variations are deemed within the scope of the present invention.

Claims (27)

THE CLAIMS DEFINING THE INVENTION ARE AS FOLLOWS:

1. A method for rated testing of a membrane module for resistance to cyclic stresses during operation of an industrial membrane separation process comprising loading at least one membrane module into a test cell of a test plant; conducting said rial membrane separation process while ting said membrane module to cyclic stresses during a plurality of membrane operating cycles, each cycle being of significantly lesser duration than used for operation of said industrial membrane separation process in an operating plant; and testing said membrane module for component failure caused by said cyclic stresses. 10

2. A method of claim 1 wherein said test plant forms part of, or is connected to, an operating plant enting said industrial membrane separation process.

3. A method of claim 1 or 2 wherein said testing allows a mean service life for the ne module to be predicted in a much lesser timeframe than actual service life. 15

4. A method of claim 3 wherein frequency of membrane operating cycles d during testing is significantly greater than the design frequency being the frequency at which stress cycles will be applied to the membrane module in normal industrial service.

5. A method of any one of the preceding claims wherein said membrane 20 operating cycles comprise filtration cycles; relaxation cycles; and/or pressure decay testing cycles in which the membranes of the ne modules are tested for leaks.

6. A method of claim 5 wherein said membrane modules are only subjected to filtration and relaxation cycles; or pressure decay testing cycles. 25

7. method of claim 5 wherein ne s are subjected to both filtration and relaxation cycles and pressure decay g cycles.

8. A method of claim 6 or 7 wherein filtration cycles correspond with periods when pump means forming part of the test cell are turned on to apply pressure to membrane; and tion cycles correspond with periods when said pump means forming part of the test cell are turned off.

9. A method of any one of the preceding claims wherein said testing plant includes a plurality of test cells.

10. A method of claim 9 wherein said test cells are operated in parallel with one test cell imposing different membrane operating cycles and a different cyclic stress regime on membrane modules than in another test cell. 1O

11. A method of any one of claims 5 to 10 wherein the frequency of filtration and relaxation cycles to which a membrane module is subjected during testing is accelerated to be significantly greater than the ncy of any pressure decay/pressure ity testing cycles

12. A method of any one of the preceding claims wherein said ne 15 module to be tested comprises a housing accommodating a bundle of hollow fibre membrane fibres, each hollow fibre extending n two opposed pots with the opposite ends of each hollow fibre being potted and sealed within said pots and component failure is ted when cracking of a pot occurs during testing.

13. A method of any one of the ing claims wherein a plurality or rack of 20 membrane modules is loaded into the test cell to enable simultaneous g.

14. A method of any one of the preceding claims wherein each membrane module is subjected to a testing cycle having a predetermined number of stress cycles to be imposed over a predetermined period of time, said predetermined number of stress cycles corresponding with required minimum design service life. 25

15. A method of claim 14 wherein said minimum design service life is measured in years and said predetermined number of stress cycles is imposed in a much shorter period in the order of weeks or months.

16. A method of any one of claims 5 to 15 wherein duration of a filtration cycle is significantly shorter than during normal service.

17. A method of claim 16 wherein said significantly shorter tion cycle duration increases filtration flux over normal service.

18. A method of claim 17 wherein filtration flux during testing is measured and compared with tion flux during normal e

19. A method of any one of claims 5 to 18 wherein testing is used to assess whether ance to membrane fouling is improved by shorter filtration cycle duration. 1O

20. A method of any one of the preceding claims including imposing chemical treatments on the membrane module at a plurality of solution strengths for varying duration.

21. A method of claim 20 wherein said chemical treatments are d during membrane operating cycles having duration the same as, or r to, 15 duration in said operating plant.

22. A method of any one of the preceding claims wherein said industrial membrane separation process requires a membrane service life of duration of a plurality of years.

23. A method of claim 22 wherein said industrial membrane separation 20 process is a liquid treatment process, preferably a water treatment process.

24. A method of claim 23 wherein said test cell is ured to simulate a membrane bioreactor with said membrane module being submerged in mixed liquor within said test cell.

25. A method of claim 23 or 24 n said water treatment process is a waste water treatment process where a membrane module will be subjected to an operating temperature greater than about 25°C, or even greater than 30°C.

26. A method of any one of claims 23 to 25 where said membrane module comprises pots made of r al and said membrane module operating temperature is greater than the glass transition temperature of said polymer material.

27. A method of any one of claims 23 to 26 wherein said testing method is applied at different temperatures to assess dependence of membrane module 10 service life on service operating temperature. L GIPPSLAND REGION WATER CORPORATION WATERMARK PATENT AND TRADE MARKS ATTORNEYS P35095NZPC Biogas Water Biogas Water