Real-time process control for an immersed membrane filtration system using a control hierarchy of discrete-state parameter changes

Abstract

An immersed membrane system or process may use measured or calculated process information to optimize one or more process operating parameters to improve performance or reduce operating costs. An on-line process control system or method may use the resistance in series method in operating an immersed membrane water treatment system. A process control system or process may consider resistance values and adjust operational parameters such as membrane aeration frequency factor, membrane aeration flow, permeate flux, permeation duration, backwash flow and duration, relaxation duration or maintenance or recovery chemical cleaning frequencies in order to reduce the operational costs related to membrane fouling removal.

Claims

1. A process for operating an immersed membrane filtration system, the immersed membrane filtration system comprising a plurality of membrane cassettes divided into trains, each train containing a sub-set of the membrane cassettes connected together for permeate removal, and a control system, the process comprising a step of: altering the operation of the system considering real-time process information according to a hierarchy including the following parameters: a. the relative duration of a higher flow rate and a lower flow rate of a gas that produces bubbles from under or between the membranes; and one or both of, b. the on and off production times for an individual train, the selectable off production times including times of at least one hour in duration; and, c. the number of trains in operation, wherein the selectable number of trains in operation includes (a) all trains in the system or (b) at least one but less than all trains in the system, wherein one parameter in the hierarchy is changed to a different discrete state before changing another parameter in the hierarchy.

2. The process of claim 1 wherein the process information includes one or more of: a. permeate flow b. reject or concentrate flowrate c. sludge filterability d. fluid temperature e. fluid viscosity f. fluid suspended solids concentration g. coagulant or other pretreatment chemical addition rate h. activated carbon dosage rate or concentration i. blower inlet air temperature j. recovery rate (permeate produced/feedwater provided) k. dissolved oxygen concentration l. oxygen uptake rate m. sludge retention time n. mixed liquor recirculation rates.

3. The process of claim 1 wherein the control system comprises a programmed controller that operates the immersed membrane filtration system at a fixed aeration rate for a period of time to establish baseline process operating conditions including permeate flow and transmembrane pressure, then automatically conducts a test in which it adjusts the aeration rate upward or downward by a certain percentage and then determines the effect of the aeration rate change on the baseline process operating conditions, then the control system assesses if the new aeration rate yields better performance in terms of either actual flow rate or flowrate per unit energy input, and based on the results of the test, the control system chooses another aeration rate to be evaluated, then this test is repeated until an optimum aeration rate has been determined.

4. The process of claim 1 wherein the control system comprises a programmed controller that conducts a test to determine the rate of increase of transmembrane pressure with time without membrane aeration occurring, the test including an initial aeration without permeation step followed by a permeation without aeration step during which transmembrane pressure is monitored, then a second aeration without permeation step follows during which the majority of the material deposited on the membrane during the earlier step is removed, then the control system correlates the transmembrane pressure increase with time against a previously input model to determine an optimum aeration rate and operating strategy.

5. The process of claim 1 wherein the process information is used to calculate a resistance value.

6. The process of claim 1 wherein the hierarchy includes: the number of trains in operation.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 is a schematic representation of an immersed membrane treatment system with a control system according to an embodiment of the invention or adapted to carry out a process according to the invention.

(2) FIG. 2 is a decision tree for a control system.

(3) FIG. 3 is another decision tree for a control system.

(4) FIG. 4 is a graph of experimental results.

(5) FIG. 5 is a schematic plan view of a membrane tank showing the arrangement of aerators.

(6) FIG. 6 is a schematic drawing of parts of an air supply system.

DETAILED DESCRIPTION

(7) Various apparatuses or processes will be described below to provide an example of an embodiment of each claimed invention. No embodiment described below limits any claimed invention and any claimed invention may cover processes or apparatuses that are not described below. The claimed inventions are not limited to apparatuses or processes having all of the features of any one apparatus or process described below or to features common to multiple or all of the apparatuses described below. It is possible that an apparatus or process described below is not an embodiment of any claimed invention. The applicants, inventors and owners reserve all rights in any invention disclosed in an apparatus or process described below that is not claimed in this document and do not abandon, disclaim or dedicate to the public any such invention by its disclosure in this document.

(8) Referring to FIG. 1, a system 8 includes an immersed membrane water treatment system 10 which is connected to one or more sensors 12 which provide input information to a computer 14 and optionally to an operator through an operator interface 16. The specific process inputs provided by the sensors 12 can include one or more of the following:

(9) Permeate flow

(10) Reject or concentrate flowrate

(11) Sludge filterability

(12) Fluid temperature

(13) Fluid viscosity

(14) Fluid suspended solids concentration

(15) Coagulant or other pretreatment chemical addition rate

(16) Activated carbon dosage rate or concentration

(17) Blower inlet air temperature

(18) Recovery rate (permeate produced/feedwater provided)

(19) Dissolved oxygen concentration (wastewater systems)

(20) Oxygen uptake rate (wastewater systems)

(21) Solids retention time (wastewater systems)

(22) Mixed liquor recirculation rates (wastewater systems)

(23) After considering the information from the sensors 12, the computer 14, after running its own programs or as instructed by an operator through the interface 16, instructs process controllers 18 to adjust the aeration regime of the system 10.

(24) The methods or step which the process controllers 18 may program may be one or more of: a. Adjustments to the airflow rate (continuous, cyclic or intermittent) Example: reduce instantaneous airflow from 0.35 m.sup.3/h-m.sup.2 to 0.30 m.sup.3/h-m.sup.2 b. Adjustments to the on and off aeration times during cyclic or intermittent aeration cycle, Examples: change cyclic aeration times from “10 seconds on/10 seconds off” to “10 seconds on/20 seconds off”; or change intermittent aeration times from “1 minute every 30 minutes” to “1 minute every 45 minutes” c. Adjustments to the on and off production times for an individual train, or Example: change from “18 hours on/6 hours off” to “3 hours on/1 hour off” d. Adjustments to the number of trains in operation or design permeate flow for each train as part of a complete system. Example: change from 2 systems operating at a higher flux to 3 systems operating at a lower flux (for example when filterability of sludge changes)

(25) Methods by which the sensors 12 may be used to collect real-time process measurements that can be used to determine changes in the aeration regime may include one or more of the following:

(26) A method may involve steps of testing, evaluating results and then optimizing. For example, process controllers 18 may operate to provide a fixed aeration rate for a period of time to establish baseline process operating conditions (including flux, transmembrane pressure). This baseline data may be conducted, for example over 4 to 8 filtration cycles, approximately 2 hrs. The system 8 would then automatically conduct a test on one or more of the trains. In this test, it would maintain the same permeate production flow but adjust the aeration rate upward or downward by a certain percentage, e.g. 5%, for as many filtration cycles as necessary to obtain steady state performance. The system 8 would monitor the process conditions (in particular TMP and fouling rate) and would be able to determine when steady state conditions are achieved. Since the development of the new boundary layer is expected to be extremely fast, steady state is expected to be 2 or 3 filtration cycles. The specific process conditions will be evaluated by the system 8 to determine if the change has affected aeration efficiency. Knowing the efficiency of the blowers and the permeate pumps, the system 8 can determine if the new aeration rate yields better performance in terms of either simple permeate flow or total energy efficiency (flowrate per unit energy input). If a decrease in blower output does not change the permeation TMP, it can be concluded that the aeration rate was in excess of that necessary to remove the suspended solids loaded on to the membrane. If a decrease in aeration rate results in an increase in TMP, then the control system can quantify the changes in operating costs and assess if the change was positive or negative. Once the analysis of the results is complete, the system can conduct a similar test at a new aeration rate. This test may be repeated until the optimum aeration rate has been determined.

(27) The system would continue to operate at the set aeration rate until changes in process operating conditions were noted e.g. an increase in the permeate flow setpoint would result in the optimization tests being repeated. Optimization tests could also be triggered based on a time interval (e.g. every 6 hours or once a day) or whenever initiated by the operator (e.g. after sludge wasting has occurred).

(28) In a method using an on-line filterability test, a train of membrane filters is operated under controlled conditions for a short period of time to test the filterability of the fluid being filtered. The test consists of three steps: 1. Stop membrane permeation but continue aeration for a period of 30 seconds to 5 minutes. The purpose of this step is to deconcentrate the liquid surrounding the membranes and to reduce the boundary layer thickness to a baseline level. 2. Stop aeration and start permeation of the membrane filters at a specified membrane flux for a period of 30 seconds to 20 minutes. During this time, the transmembrane pressure will rise due to the redevelopment of the boundary layer at the surface of the membrane filters. The purpose of this step is to develop a relationship between transmembrane pressure and time. This relationship can be correlated to an optimum aeration rate. 3. Stop membrane permeation and resume aeration for a period of 30 seconds to 5 minutes. The purpose is to remove the solids that accumulated on the membranes during step 2. 4. Based on the results of the test, and comparing the results with previously input model results, adjust aeration rate or operating strategy (refer to methods a) to d) described above and the train is ready to operate under new optimized conditions.

(29) In a process parameter modeling method, the control system would use real-time data to determine the membrane system operating conditions and based on those conditions and previously input models, would set the aeration flow or system operating strategy accordingly. The optimization can be as simple as tracking a single parameter (e.g. as permeate flow increases, so does aeration rate) or as many parameters as is necessary can be used. For example, in wastewater treatment systems, the optimum aeration rate could be a function of the sludge filterability and permeate flowrate. A model is first developed and the data incorporated into the system 8. The models necessary will depend on each application and the parameters that can be expected to change with operation. Another example would be if the sludge filterability is good, 2 trains can be operated at a higher flow and when the sludge filterability is low, three trains can be operated at proportionately lower flows.

(30) A method may use measured or calculated resistance data, for example one or more of the resistance parameters in the resistance in series model. The resistance in series model represents the total resistance to filtration as the combination of a number of independent resistances. Resistance values may be used in, for example, feedback, feed forward, adaptive or model predictive control processes.
R.sub.t=R.sub.m+R.sub.a+R.sub.b+R.sub.c

(31) R.sub.t is the total resistance, m.sup.−1

(32) R.sub.m is the membrane resistance, m.sup.−1 as represented by the membrane clean water test

(33) R.sub.a is the adsorptive fouling resistance, caused by fouling agents adsorbed on the surface and in the porous structure of the membrane.

(34) R.sub.b is the pore blocking resistance, m.sup.−1, caused by colloidal matter and micro particles that are comparable in size to the membrane pore size. This is normally represented by a rapid rise in transmembrane pressure immediately following a backwash.

(35) R.sub.c is the cake resistance, m.sup.−1 which progressively increases between two backwash cycles as the cake builds up on the membrane surface.

(36) For this particular application, the value of R.sub.m is considered as constant and can be obtained from a clean water test. The resistance during backwash (R.sub.db) will be used to calculate the adsorptive fouling resistance R.sub.a and the membrane resistance as follows:
R.sub.a=R.sub.db*f−R.sub.m

(37) Where f is a factor that relates resistance during permeation to the resistance during backwash.

(38) The cake resistance (R.sub.c) will be estimated by performing a linear regression for those experimental data of a given permeation cycle that corresponds to the permeation mode and comply with one or more of the following conditions, called Valid Permeation Cycle (VPC) conditions.
J=J.sub.setpoint±35%;

(39) where J.sub.setpoint is the permeate pump set point. Alternate or additional criteria may also be used for VPC conditions. For example, the VPC conditions may be that J=J.sub.setpoint±35% and dR.sub.t/dT<5% for 5 consecutive samples. dR.sub.t/dT is the rate of increase of total resistance in time during permeation and it is calculated as:
dR.sub.t/dT=(R.sub.n+1−R.sub.n)/(T.sub.n+1−T.sub.n)

(40) Where R.sub.n and R.sub.n+1 are two consecutive total resistance data points of any given permeation cycle and T.sub.n and T.sub.n+1 are the corresponding permeation times.

(41) Then the cake resistance will be calculated using the experimental data that complies with the above illustrated conditions.
R.sub.c=M(T.sub.2−T.sub.1)

(42) Where M is the slope of the linear regression; T.sub.1 and T.sub.2 are the starting and ending times of the permeation cycle; respectively.

(43) The pore blocking resistance will be estimated as follows:
R.sub.b=R.sub.ab*e−R.sub.m

(44) Where e is the cake removal efficiency and represents the amount of cake remaining on the membrane surface after the application of relaxation or backwash and is a function of the cake stickiness and the operational conditions of the relaxation or backwash. R.sub.ab is the resistance after backwash and is determined by the average of the first five samples of a permeation cycle that meet the VPC condition as outlined in the cake resistance calculation.
R.sub.ab=(R.sub.t1+R.sub.t2+R.sub.t3R.sub.t4+R.sub.t5)/5

(45) The first step in the control strategy involves the calculation of the filtration resistances during operation at permeation mode using on-line MBR operational data, in a similar fashion as described above.

(46) The calculation of the resistance after backwash (R.sub.ab), resistance after backwash increase rate (ΔR.sub.b) and cake filtration (R.sub.e) resistances is of particular interest for the adjustment of the value of the membrane aeration frequency factor, membrane aeration flow, relaxation cycle duration, permeation cycle duration, backwash flow, backwash duration to reduce the membrane fouling rate, using an on-line process control. The calculation procedure is as follows:

(47) Measure TMP, Permeate Flux, Temperature and Time for two consecutive permeation and backwash or relaxation cycles.

(48) Calculate ΔR.sub.ab and R.sub.c for any given permeation cycle as follows:
ΔR.sub.ab=R.sub.ab(cycle 2)−R.sub.ab(cycle 1)  i.
R.sub.c=M(T.sub.2−T.sub.1)  ii.

(49) Compare ΔR.sub.ab and R.sub.c values with their corresponding set points (Table 1.0) to adjust the value of the membrane aeration frequency factor, membrane aeration flow, relaxation cycle duration, permeation cycle duration, backwash flow and duration, maintenance and recovery clean frequencies so as to minimize the energy required for membrane fouling removal. The corresponding operational changes may be performed every cycle, if needed.

(50) The calculation of the adsorptive resistance (R.sub.a) is optional but may be of particular interest for the adjustment of the value of the maintenance and recovery chemical cleaning frequencies using an on-line process control, if desired. However, calculation of R.sub.a may also be omitted if chemical cleaning procedures will not be controlled. The calculation procedure is as follows:

(51) 1. Initiate a filtration cycle (1) followed by backwash period followed by another filtration cycle (2). In those cases where relaxation is used as the mechanism for cake removal it would be required to switch to backpulse mode after a number of cycles in order to collect the information needed to estimate the membrane condition.

(52) 2. Measure TMP, Permeate Flux, Temperature and Time for two consecutive filtration and backwash cycles.

(53) 3. Calculate (R.sub.a) as follows:
(R.sub.a)=resistance during backwash (cycle 2)−resistance during backwash (cycle 1)

(54) 4. Compare (R.sub.a) values with its corresponding set point to adjust the value of the maintenance and recovery chemical cleaning frequencies as to minimize the energy required for membrane fouling removal. The corresponding operational changes will be performed after five cycles, if needed.

(55) Two different groups of set points may be established for a particular system; sustainable and optimized. The values of these set points might change for different treatment plants as they take into account different operational variables (e.g. mixed liquor characteristics, wastewater temperature) that are known to be site related. The set points may be determined during piloting of a system or based on historical system performance. One or more of the set points may also vary with time. The different resistances that are being monitored, for example ΔR.sub.ab and R.sub.c, will be compared against their respective set points in order to perform the adequate operational changes. The calculated resistances may be single values or a composite of several values spaced in time, for example as obtained by a mathematical averaging or regression.

(56) The sustainable set point represents the maximum value of resistance at which the system should be operated. When any of the values of the resistances being monitored are higher than any of the corresponding sustainable set points (Red Zone); only those operational changes will be made that ensure that a reduction in the membrane fouling is achieved (Fouling Removal mode).

(57) On the other hand, the optimized set point is the maximum value for which operational changes to achieve energy savings are possible; when any of the values of the resistances being monitored are between the sustainable and the optimized set point (Yellow Zone), no change of the operational parameters will be performed. When all of the values of the resistances being monitored are lower than the values of the corresponding optimized set point (Green Zone), then it is possible to execute operational changes that lead to energy savings (Energy Savings mode).

(58) If one of the set points is exceeded, then the system will be considered to be operating at the zone corresponding to that set point. The two set points enhance the stability of the process, help avoid switching system operations too frequently or in response to errant resistance measurements and allow for operation for extended periods of time within a range between the two set points. However, a single set point can be used if the process is otherwise dampened. For example, more robust regression algorithms can be used on the calculated resistances, a mathematical band may be constructed around the set point, the sampling rate may be decreased, the calculated resistances may be required to be above or below the set point at multiple sampling periods or other techniques or combinations of techniques can be used to dampen the system.

(59) FIGS. 2 and 3 show decision trees for systems having a single parameter, A.F.F., adjustable between two operational states. The two states are 10 seconds on 10 seconds off and 10 seconds on 30 seconds off. In the 10/10 cycle, a number of air blowers are used and the total air supply is cycled between two halves of a number of aerators. In the 10/30 cycle, half of the air blowers are turned off and the total air is cycled between four quarters of the number of aerators. For example, FIG. 5 shows a membrane tank 30 having several conduit aerators 32 each connected to one of two tank manifolds 34 running along the bottom of the tank 30. Tank manifold 34a delivers air to one half of aerators 32 while tank manifold 34b delivers air to the other half of aerators 32. Drop legs 36 extend upwards from tank manifolds 34 and allow air to be fed from above the tank 30 to the grid of aerators 32. Referring to FIG. 6, 4 tanks 30 each contain a number of immersed membrane cassettes 38 placed over the grid of aerators 32. The cassettes 38 are connected together for permeate removal and may be called a train. Aerators 32, tank manifolds 34 and drop legs 36 are not shown in FIG. 6 but are also present in each tank 30 of FIG. 6 as shown in FIG. 5. The drop legs 36 from each tank 30 connect to connectors 40 of supply system 42. Supply system 42 further includes a set of valves 44 in pipes 46 between connectors 40 and a plant manifold 48. Plant manifold 48 is connected to four blowers 50. Valves 44 are powered by solenoids of fluidic actuators and connected to process controllers 18. During 10/10 aeration, all four blowers 50 are on, and valves 44 are controlled such that valves 44a,c,e,g are open for 10 seconds while valves 44b,d,f,h are closed then valves 44a,c,e,g are closed for 10 seconds while valves 44b,d,f,h are open. This cycle is repeated for as long as the 10/10 cycle is requested. Thus, each cassette always has a flow of air to it, but the flow switches between aerators 32 connected to the different tank manifolds 34. When a 10/30 cycle is requested, two trains are linked such that air flow is alternated between the two tank manifolds 34 in each tank 30 and between the two tanks 30, that is between four tank manifolds 34 in two tanks 30. Thus each cassette 38 experiences a 10 seconds on 10 seconds off aeration pattern but with the source of air in consecutive air on periods alternating between the two sets of aerators 32 in that tank 30. Each aerator 32 has a 10 seconds on 30 seconds off pattern. In greater detail, for 10 seconds valves 44a,e are open while the others are closed. For the next 10 seconds, valves 44c,g are open while the others are closed. For the next 10 seconds, valves 44b,f are open while the others are closed. For the next 10 seconds, valves 44d,h are open while the others are closed. This pattern is repeated for as long as 10/30 aerations is requested. Other methods of reducing a time and space averaged air flow without reducing instantaneous flow to individual aerators might also be used. For example, in a system where air is supplied continuously or cyclically or intermittently from a set of blowers to a set of aerators, half of the blowers may be turned off and a valve closed to isolate half of the aerators. Although the discussion above describes two or three types of resistance values, one, two, three or more than three resistance values may be used. For example, in the system of FIG. 2, cake resistance may be the only parameter used since it is closely related to air scouring.

EXAMPLES

Example 1

(60) In a 2 month test period the application of an on-line MBR process control, based on the results from the resistance in series method was studied. This on-line MBR process control will adjust different operational parameters (e.g. membrane aeration frequency factor, relaxation duration, etc.) to reduce the MBR operational costs or increase membrane fouling removal, as required.

(61) A ZeeWeed® immersed membrane pilot plant, made by ZENON Environmental Inc, was operated using raw wastewater feed from a municipal water pollution control centre. The raw wastewater was screened through a 0.75 mm screen. The pilots were operated at a hydraulic retention time of 6 hours and a sludge retention time of 15 days.

(62) At the first set of conditions that lasted 2 days, ZeeWeed® membranes were operated for a 10 minute permeation cycle with a net flux of 14 gallons/(square foot*day) (gfd), a corresponding instantaneous flux of 15.4 gfd and 1 minute relaxation time. The systems were operated at a mixed liquor suspended solids concentration of around 10 g/l.

(63) The membrane aeration frequency factor was set at 0.25 (10 seconds on and 30 second off) during permeation and relaxation; a coarse air bubble flow rate of 8 scfm per gap was used. The bioreactor was aerated using fine bubble aerators. Mixed liquor was recirculated from the bioreactor to the membrane tank by a pump and was returned to the bioreactor by gravity. Sludge was wasted on intermittent basis to maintain a steady sludge retention time.

(64) A detailed analysis of the characteristics of this MBR system allowed for the identification of the corresponding on-line MBR process control set points for this system; these are presented in Table 1.0. The yellow zone is an operation zone between the red zone and the green zone although it has no distinct operation mode name or parameters.

(65) TABLE-US-00001 TABLE 1.0 Set points of the MBR on-line process control. Operation Parameter zone Set point Operation mode ΔR.sub.ab R.sub.c R.sub.a Red Sustainable Fouling Removal 2*10.sup.8 8*10.sup.11 2*10.sup.8 Green Optimized Energy Savings 1*10.sup.8 6*10.sup.11 1*10.sup.8

(66) Tables 1.1 and 1.2 contain some of the operational data corresponding to two consecutive permeation cycles of operation under the above described operational conditions. These operational data will be used to describe the resistance calculations.

(67) TABLE-US-00002 TABLE 1.1 Permeation cycle 1 experimental data (experimental conditions 1). FLUX TMP Temp Rt Time (gfd) (psi) (° C.) μ (m.sup.−1) dR/dT 08:50:41 0.0 0.04 12.73 0.0 N/A N/A 08:50:46 0.0 0.02 12.73 0.0 N/A N/A 08:50:51 0.0 0.03 12.72 0.0 N/A N/A 08:50:56 0.0 0.03 12.72 0.0 N/A N/A 08:51:01 0.0 0.03 12.68 0.0 N/A N/A 08:51:07 0.0 0.03 12.69 0.0 N/A N/A 08:51:12 0.0 0.03 12.69 0.0 N/A N/A 08:51:17 0.0 0.01 12.70 0.0 N/A N/A 08:51:22 0.0 −0.07 12.71 0.0 N/A N/A 08:51:27 0.0 0.02 12.72 0.0 N/A N/A 08:51:32 0.0 −0.02 12.74 0.0 N/A N/A 08:51:37 0.0 0.84 12.73 0.0 N/A N/A 08:51:42 8 1.39 12.72 0.0 2.57E+12 6 08:51:47 15.5 2.90 12.70 0.0 2.73E+12 2 08:51:52 15.6 2.96 12.71 0.0 2.79E+12 0 08:51:57 15.5 2.99 12.70 0.0 2.79E+12 1 08:52:02 15.7 2.99 12.69 0.0 2.81E+12 1 08:52:07 15.6 3.01 12.68 0.0 2.83E+12 0 08:52:12 15.7 3.02 12.70 0.0 2.83E+12 −1 08:52:17 15.7 3.03 12.71 0.0 2.81E+12 0 08:52:22 15.4 3.03 12.72 0.0 2.79E+12 1 08:52:27 15.5 3.00 12.71 0.0 2.83E+12 0

(68) TABLE-US-00003 TABLE 1.2 Permeation cycle 2 experimental data (experimental conditions 1). FLUX TMP Temp Rt Time (gfd) (psi) (° C.) μ (m.sup.−1) dR/dT 09:01:42 0.000 0.09 12.65 0.001 N/A N/A 09:01:47 0.000 0.06 12.64 0.001 N/A N/A 09:01:52 0.000 0.04 12.63 0.001 N/A N/A 09:01:57 0.000 0.05 12.62 0.001 N/A N/A 09:02:02 0.000 0.04 12.63 0.001 N/A N/A 09:02:07 0.000 0.04 12.63 0.001 N/A N/A 09:02:12 0.000 0.16 12.64 0.001 N/A N/A 09:02:17 0.000 0.03 12.65 0.001 N/A N/A 09:02:22 0.000 −0.01 12.66 0.001 N/A N/A 09:02:27 0.000 0.02 12.65 0.001 N/A N/A 09:02:32 0.000 0.02 12.65 0.001 N/A N/A 09:02:37 0.000 0.02 12.64 0.001 N/A N/A 09:02:42 4.753 0.77 12.64 0.001 2.37E+12 15 09:02:48 9.888 1.89 12.64 0.001 2.80E+12 1 09:02:53 15.777 3.03 12.63 0.001 2.83E+12 2 09:02:58 15.550 3.08 12.64 0.001 2.90E+12 3 09:03:03 15.410 3.20 12.64 0.001 2.97E+12 0 09:03:08 15.498 3.09 12.65 0.001 2.98E+12 −5 09:03:13 15.593 3.09 12.66 0.001 2.84E+12 2 09:03:18 15.601 3.06 12.66 0.001 2.89E+12 0 09:03:23 15.667 3.10 12.65 0.001 2.89E+12 1 09:03:28 15.550 3.12 12.64 0.001 2.93E+12 0 09:03:33 15.615 3.13 12.64 0.001 2.93E+12 2 09:03:38 15.396 3.13 12.64 0.001 2.99E+12 0

(69) The operational data corresponding to these conditions was used to calculate the different relevant resistances.

(70) As mentioned above, the resistance after backwash will be calculated based on the average of the first five values that comply with the valid permeation cycle conditions. The adsorptive fouling resistance was not calculated in the following examples.

(71) As it can be observed from the presented table, for the first permeation cycle, the values prior to 8:51:47 do not meet the VPC conditions and consequently cannot be used for this calculation. The values corresponding from 8:51:47 to 8:52:07 are the first five values that comply with these conditions. For the second permeation cycle the same procedure is used. The value of R.sub.ab is calculated as follows:
R.sub.ab(cycle 1)=(2.73*10.sup.12+2*2.79 10.sup.12+2.81*10.sup.12+2.83*10.sup.12)/5=2.79*10.sup.12 m.sup.−1
R.sub.ab(cycle 2)=(2.80*10.sup.12+2.83 10.sup.12+2.90*10.sup.12+2.97*10.sup.12+2.98*10.sup.12)/5=2.89*10.sup.12 m.sup.−1
ΔR.sub.ab=R.sub.ab(cycle 2)−R.sub.ab(cycle 1)=2.89*10.sup.12−2.79*10.sup.12=1.0*10.sup.11 m.sup.−1

(72) As it has been previously established, R.sub.c can be estimated as follows:
R.sub.c=M(T.sub.2−T.sub.1)

(73) After performing the linear regression the value of the slope M was determined:
M=1.18*10.sup.11 m.sup.−1/min
R.sub.c=1.18*10.sup.11 m.sup.−1/min*(9.99 min)=1.17*10.sup.12 m.sup.−1

(74) From the analysis of the calculated resistances it can be concluded that the values of ΔR.sub.ab and R.sub.c are both higher than the corresponding sustainable set points, which means that the system is operating at the Red Zone and the application of the Fouling Removal mode is needed to avoid any slugging of the membrane modules.

(75) Following the established control hierarchy for the Fouling Removal mode, the aeration frequency factor was increased from 0.25 (10 on/30 off) to 0.5 (10 on/10 off) while maintaining the remaining operational parameters constant. Immediately after the change in aeration frequency factor, the operational data corresponding to these new operational conditions was used to calculate the different relevant resistances to assess the effectiveness of these measures on fouling removal. The procedure is the same as done previously.

(76) Tables 1.3 and 1.4 contain some of the operational data corresponding to two consecutive permeation cycles performed on the Energy Savings mode using an aeration frequency factor of 0.25. These operational data will be used to describe the resistance calculations.

(77) TABLE-US-00004 TABLE 1.3 Permeation cycle 1 experimental data (experimental conditions 2). FLUX TMP Temp Rt Time (gfd) (psi) (° C.) μ (m.sup.−1) dR/dT 00:05:42 0 0.016 14.256 0.001 N/A N/A 00:05:47 0 −0.109 14.260 0.001 N/A N/A 00:05:52 0 0.003 14.244 0.001 N/A N/A 00:05:57 0 0.011 14.240 0.001 N/A N/A 00:06:02 0 0.019 14.214 0.001 N/A N/A 00:06:07 0 −0.109 14.206 0.001 N/A N/A 00:06:12 0 0.003 14.191 0.001 N/A N/A 00:06:17 0 0.011 14.176 0.001 N/A N/A 00:06:22 0 −0.006 14.160 0.001 N/A N/A 00:06:27 0 −0.029 14.149 0.001 N/A N/A 00:06:32 0 −0.001 14.160 0.001 N/A N/A 00:06:37 0 0.005 14.153 0.001 N/A N/A 00:06:42 6.350 0.850 14.157 0.001 1.940E+12 −1.68 00:06:47 14.466 1.886 14.157 0.001 1.908E+12 −3.77 00:06:52 15.117 1.900 14.160 0.001 1.839E+12 0.29 00:06:57 15.154 1.910 14.153 0.001 1.844E+12 −0.46 00:07:02 15.264 1.920 14.153 0.001 1.836E+12 1.37 00:07:07 15.117 1.921 14.153 0.001 1.861E+12 −0.01 00:07:12 15.147 1.929 14.157 0.001 1.861E+12 −0.01 00:07:17 15.213 1.930 14.179 0.001 1.861E+12 −1.11 00:07:22 15.271 1.919 14.199 0.001 1.840E+12 1.21 00:07:28 15.227 1.936 14.210 0.001 1.863E+12 −0.78

(78) TABLE-US-00005 TABLE 1.4 Permeation cycle 2 experimental data (experimental conditions 2). FLUX TMP Temp Rt Time (gfd) (psi) (° C.) μ (m.sup.−1) dR/dT 00:16:47 0.000 0.011 14.286 0.001 N/A N/A 00:16:52 0.000 −0.037 14.275 0.001 N/A N/A 00:16:57 0.000 0.007 14.267 0.001 N/A N/A 00:17:02 0.000 0.035 14.256 0.001 N/A N/A 00:17:07 0.000 0.008 14.233 0.001 N/A N/A 00:17:12 0.000 0.042 14.221 0.001 N/A N/A 00:17:17 0.000 0.004 14.210 0.001 N/A N/A 00:17:22 0.000 −0.023 14.199 0.001 N/A N/A 00:17:27 0.000 0.002 14.172 0.001 N/A N/A 00:17:32 0.000 −0.011 14.172 0.001 N/A N/A 00:17:37 0.000 0.007 14.172 0.001 N/A N/A 00:17:42 0.000 0.777 14.153 0.001 N/A N/A 00:17:47 7.398 1.020 14.160 0.001 1.987E+12 −7.99 00:17:52 15.073 1.893 14.172 0.001 1.840E+12 0.75 00:17:57 15.073 1.912 14.160 0.001 1.854E+12 −0.17 00:18:03 15.198 1.915 14.137 0.001 1.851E+12 0.57 00:18:08 15.183 1.931 14.141 0.001 1.861E+12 0.04 00:18:13 15.154 1.931 14.160 0.001 1.862E+12 −0.43 00:18:18 15.235 1.930 14.160 0.001 1.854E+12 0.29 00:18:23 15.249 1.935 14.179 0.001 1.859E+12 −0.12 00:18:28 15.227 1.939 14.191 0.001 1.857E+12 −1.32 00:18:33 15.396 1.933 14.199 0.001 1.833E+12 −0.26 00:18:38 15.344 1.911 14.202 0.001 1.828E+12 1.90 00:18:43 15.161 1.929 14.206 0.001 1.864E+12 −0.25

(79) The operational data corresponding to these new conditions was used to calculate the different relevant resistances.

(80) As mentioned before, the resistance after backwash for both permeation cycles will be calculated based on the average of the first five values that comply with the valid permeation cycle conditions. As it can be observed from the above table; the values prior to 0:06:47 do not meet the VPC conditions and consequently cannot be used for this calculation. The values corresponding from 0:06:47 to 0:07:07 are the first five values that comply with these conditions. For the second permeation cycle the same procedure is used. The value of R.sub.ab is calculated as follows:
R.sub.ab(cycle 1)=(1.908*10.sup.12+1.839*10.sup.12+1.844*10.sup.12+1.836*10.sup.12+1.861*10.sup.12)/5=1.857*10.sup.12 m.sup.−1
R.sub.ab(cycle 2)=(1.840*10.sup.12+1.854*10.sup.12+1.851*10.sup.12+1.861*10.sup.12+1.862*10.sup.12)/5=1.853*10.sup.12 m.sup.−1
ΔR.sub.ab=R.sub.ab(cycle 2)−R.sub.ab(cycle 1)=1.853*10.sup.12−1.857*10.sup.12=−4*10.sup.9 m.sup.−1

(81) As it has been previously established, R.sub.c can be estimated as follows:
R.sub.c=M(T.sub.2−T.sub.1)

(82) After performing the linear regression the value of the slope M was determined:
M=1.65*10.sup.10 m.sup.−1/min
R.sub.c=1.65*10.sup.10 m.sup.−1/min*(9.99 min)=1.648*10.sup.11 m.sup.−1

(83) From the analysis of the calculated resistances it can be concluded that the values of ΔR.sub.ab and R.sub.c are both lower than their corresponding optimized set points, which means that the system is operating at the Green Zone and the application of the Energy Savings mode is available to reduce the MBR operational costs.

(84) Following the established control hierarchy for the Energy Savings mode, the relaxation period duration was decreased from 1 min to 30 seconds and the net permeate flux was increased from 14 to 16 gfd by increasing the instantaneous flux from 15.4 to 17.6 gfd, while maintaining the remaining operational parameters constant. Optionally, the net permeate flux alone could have been adjusted as was done in the long term tests described further below. Although the control hierarchy gives a higher preference to other operational changes such as the decrease of the aeration frequency factor or aeration flow and the increase of the permeation cycle duration, these changes were difficult to implement due to intrinsic limitations of the MBR system in this example.

(85) Immediately after the change in relaxation period duration and increase in net permeate flux, the operational data corresponding to these new operational conditions was used to calculate the different relevant resistances to assess the effectiveness of these measures on maintaining stable operational conditions while allowing for MBR operational costs savings. The procedure is the same as done previously.

(86) Tables 1.5 and 1.6 contain some of the operational data corresponding to two consecutive permeation cycles of operation under the above described operational conditions. These operational data will be used to describe the resistance calculations.

(87) TABLE-US-00006 TABLE 1.5 Permeation cycle 1 experimental data (experimental conditions 3). FLUX TMP Temp Rt Time (gfd) (psi) (° C.) μ (m.sup.−1) dR/dT 00:09:45 0.000 0.068 17.803 0.001 N/A N/A 00:09:50 0.000 −0.070 17.800 0.001 N/A N/A 00:09:55 0.000 0.013 17.800 0.001 N/A N/A 00:10:00 0.000 0.020 17.800 0.001 N/A N/A 00:10:05 0.000 0.044 17.796 0.001 N/A N/A 00:10:10 0.000 −0.112 17.800 0.001 N/A N/A 00:10:15 6.607 0.808 17.788 0.001 1.984E+12 4.781757 00:10:20 15.564 2.227 17.784 0.001 2.084E+12 −7.69263 00:10:25 16.956 2.284 17.765 0.001 1.935E+12 1.196952 00:10:30 16.948 2.270 17.769 0.001 1.959E+12 1.201163 00:10:35 16.846 2.282 17.762 0.001 1.982E+12 −0.05997 00:10:40 16.941 2.296 17.750 0.001 1.981E+12 0.573327 00:10:45 16.941 2.306 17.750 0.001 1.993E+12 0.402593 00:10:50 16.919 2.313 17.758 0.001 2.001E+12 0.783189 00:10:56 16.890 2.323 17.769 0.001 2.017E+12 −1.33901 00:11:01 17.117 2.334 17.765 0.001 1.990E+12 0.347578 00:11:06 16.919 2.296 17.765 0.001 1.997E+12 −0.7252 00:11:11 17.088 2.301 17.777 0.001 1.982E+12 −0.55843 00:11:16 16.985 2.292 17.773 0.001 1.971E+12 −0.22469 00:11:21 17.029 2.287 17.773 0.001 1.967E+12 −0.17168 00:11:26 17.190 2.305 17.773 0.001 1.964E+12 2.277346 00:11:31 16.890 2.317 17.758 0.001 2.009E+12 0.058487

(88) TABLE-US-00007 TABLE 1.6 Permeation cycle 2 experimental data (experimental conditions 3). FLUX TMP Temp Rt Time (gfd) (psi) (° C.) μ (m.sup.−1) dR/dT 00:20:19 0.000 0.048 17.765 0.001 N/A N/A 00:20:24 0.000 0.082 17.773 0.001 N/A N/A 00:20:29 0.000 0.010 17.769 0.001 N/A N/A 00:20:34 0.000 0.021 17.773 0.001 N/A N/A 00:20:39 0.000 0.030 17.762 0.001 N/A N/A 00:20:44 0.000 0.004 17.754 0.001 N/A N/A 00:20:49 6.658 0.838 17.750 0.001 1.971E+12 7.160825 00:20:54 15.125 2.175 17.746 0.001 2.124E+12 −11.0537 00:20:59 17.139 2.259 17.731 0.001 1.912E+12 1.299441 00:21:04 17.124 2.262 17.712 0.001 1.937E+12 0.266417 00:21:09 17.190 2.286 17.708 0.001 1.943E+12 1.798625 00:21:14 17.007 2.291 17.704 0.001 1.978E+12 0.18199 00:21:19 17.022 2.315 17.704 0.001 1.982E+12 −0.7813 00:21:24 17.205 2.312 17.712 0.001 1.966E+12 1.718922 00:21:29 17.007 2.321 17.720 0.001 2.001E+12 1.289662 00:21:34 16.861 2.337 17.712 0.001 2.027E+12 −3.71094 00:21:39 16.956 2.407 17.697 0.001 1.954E+12 0.680588 00:21:44 17.139 2.313 17.712 0.001 1.968E+12 0.559187 00:21:49 16.861 2.308 17.704 0.001 1.979E+12 1.170542 00:21:54 16.802 2.299 17.712 0.001 2.002E+12 −0.89163 00:22:00 16.970 2.310 17.697 0.001 1.985E+12 1.07242 00:22:05 16.941 2.322 17.697 0.001 2.006E+12 1.359174 00:22:10 16.787 2.333 17.697 0.001 2.034E+12 −1.23236 00:22:15 17.036 2.342 17.689 0.001 2.009E+12 −0.00051

(89) The operational data corresponding to these different operational conditions was used to calculate the different relevant resistances.

(90) As mentioned before, the resistance after backwash for both permeation cycles will be calculated based on the average of the first five values that comply with the valid permeation cycle conditions. As it can be observed from the above table; for the first permeation cycle the values prior to 0:10:20 do not meet these requirements and consequently cannot be used for this calculation. The values corresponding from 0:10:20 to 0:10:40 are the first five values that comply with these conditions. For the second permeation cycle the same procedure is used. The value of R.sub.ab is calculated as follows:
R.sub.ab(cycle 1)=(2.084*10.sup.12+1.935*10.sup.12+1.959*10.sup.12+1.982*10.sup.12+1.981*10.sup.12)/5=1.988*10.sup.12 m.sup.−1
R.sub.ab(cycle 2)=(2.124*10.sup.12+1.912*10.sup.12+1.937*10.sup.12+1.943*10.sup.12+1.978*10.sup.12)/5=1.979*10.sup.12 m.sup.−1
ΔR.sub.ab=R.sub.ab(cycle 2)−R.sub.ab(cycle 1)=1.979*10.sup.12 m.sup.−1−1.988*10.sup.12=m.sup.−1=−9.00*10.sup.9 m.sup.−1

(91) As it has been previously established, R.sub.c can be estimated as follows:
R.sub.c=M(T.sub.2−T.sub.1)

(92) After performing the linear regression the value of the slope M was determined:
M=1.67*10.sup.10 m.sup.−1/min
R.sub.c=1.67*10.sup.10 m.sup.−1/min*(9.99 min)=1.66*10.sup.11 m.sup.−1

(93) From the analysis of the calculated resistances it can be concluded that the values of ΔR.sub.ab and R.sub.c are both lower than their corresponding sustainable set points, which means that the system is operating at the Green Zone and the application of the Energy Savings mode is available to reduce the MBR operational costs, while maintaining sustainable operational conditions in the system.

(94) FIG. 4 presents the long term results obtained from testing using a system similar to that above except that dR/dT was not used as a VPC condition and A.F.F. was the only controlled parameter. FIG. 3 shows the net permeate flux with the obtained cake resistance for each permeation cycle along with the values of the sustainable and optimized cake resistance, respectively. During this testing, frequent peak flux events were simulated by increasing the net permeate flux from 16 up to 26 gfd. These peak flux events were performed to assess the sensitivity of the on-line process control system when faced with operational disturbances.

(95) As it can be observed in FIG. 4, as the net permeate flux increased from 16 gfd to 26 gfd the cake resistance also increased. During some of the peak flux events the obtained values of cake resistance exceeded the sustainable cake resistance set point value causing the system to operate in Fouling Removal mode hence using an aeration frequency factor of 0.5.

(96) However, during most permeation cycles the obtained values of cake resistance were around the optimized set point allowing the system to operate in the Energy Savings mode hence using an aeration frequency factor of 0.25. The obtained values of ΔR.sub.ab were used but are not presented in FIG. 4. The observed trend of ΔR.sub.ab was very similar to that of the cake resistance. Table 1.7 presents the percentage and number of cycles performed at each aeration frequency factor.

(97) TABLE-US-00008 TABLE 1.7 On-line process control system long term testing summary. A.F.F Aeration Frequency % Performed Number of Cycles 0.25 10 ON/30 OFF 87 4007 0.5 10 ON/10 OFF 13 602

(98) As it can be observed in Table 1.7 for the vast majority of the permeation cycles the air consumption was reduced by 50% by prolonging the OFF time from 10 up to 30 seconds, which led to a significant reduction in the energy requirements of the MBR system.