Abstract
A control method is provided for a filter system, which includes at least one filter element (2). The method includes continuously recording a total energy consumption (E.sub.G) during a filtration cycle (22) of the filter system. The total energy consumption (E.sub.G) includes at least of the energy consumption (E.sub.B) for a physical cleaning (24) and the energy consumption (E.sub.P) for the subsequent production cycle (23) up to a predefined, in particular current point in time. The method further includes computing a relative energy consumption (E.sub.rel) by way of division of the recorded total energy consumption (E.sub.G) by a net permeate volume (Q.sub.N) which has been produced during the filtration cycle (22) up to the predefined point in time and starting a physical cleaning (24) in dependence on the relative energy consumption or of a characteristic value derived from this.
Claims
1. A control method for a filter system, which filter system comprises at least one filter element, the method comprising the steps of: continuously recording a total energy consumption during a filtration cycle of the filter system, wherein the total energy consumption is comprised of energy consumption for a physical cleaning and energy consumption for a subsequent production cycle up to at predefined point in time; computing a relative energy consumption by way of division of the recorded total energy consumption by a net permeate volume which has been produced during the filtration cycle up to the predefined point in time; and starting a physical cleaning in dependence on the relative energy consumption or in dependence on a characteristic value derived from the relative energy consumption.
2. A control method according to claim 1, wherein a physical cleaning is started at the current point in time, when the relative energy consumption or the characteristic value derived from the relative energy consumption has reached a predefined value.
3. A control method according to claim 1, wherein the physical cleaning is begun when a magnitude of a derivative of the continuously determined relative energy consumption has reached a predefined gradient or has passed a minimum.
4. A control method according to claim 1, wherein the net permeate volume corresponds to a produced permeate volume minus that volume which is used for a physical cleaning and/or a cleaning-in-place.
5. A control method according to claim 1, wherein the recording of the total energy consumption and of the net permeate volume is started anew before carrying out a physical cleaning or carrying out a cleaning-in-place.
6. A control method according to claim 5, wherein the physical cleaning is a backwashing, a crossflow, a mechanical scraping and/or an air scouring.
7. A control method according to claim 5, wherein at least one system parameter is continuously monitored during the physical cleaning, and the physical cleaning is carried out until this system parameter remains stable.
8. A control method according to claim 7, wherein the at least one system parameter is a hydraulic resistance of the filter system.
9. A control method according to claim 8, wherein a pressure across the filter element or filter system and a flow through the filter element are detected during the physical cleaning, for detecting the hydraulic resistance.
10. A control method according to claim 1, wherein a cleaning-in-place is started when a pressure across the filter element reaches a predefined maximum.
11. A control method according to claim 10, wherein the point in time for the next cleaning-in-place is determined by way of extrapolation of a curve which defines the pressure across the filter system or a filter element at the end of the filtration cycle, with respect to the number of filtration cycles.
12. A control method according to claim 3, wherein the predefined gradient is adapted in dependence on an estimated energy consumption which has been determined on the basis of the energy consumption of a number of preceding filtration cycles.
13. A control method according to claim 12, wherein for adapting the predefined gradient, an extrapolation of the energy consumption of a number of filtration cycles is computed, in order to predict the energy consumption of a current cleaning-in-place cycle, and that such a predicted energy consumption is compared with a previously predicted energy consumption for the current cleaning-in-place cycle.
14. A control method according to claim 13, wherein the predefined gradient is increased when the previously predicted energy consumption is smaller than the currently predicted energy consumption.
15. A control method according to claim 13, wherein the predefined gradient is reduced when the previously predicted energy consumption is greater than the currently predicted energy consumption.
16. A control method according claim 1, wherein: the continuously recorded total energy consumption corresponds at least to an energy consumption for a cleaning-in-place, to the energy consumption for the effected physical cleanings after a last cleaning-in-place and to the energy consumption for subsequent production cycles up to a predefined point in time, the relative energy consumption is continuously determined on the basis of this total energy consumption; and a cleaning-in-place is started when the relative energy consumption has reached or passed a minimum.
17. A control method according to claim 16, wherein the relative energy consumption is continuously computed and is plotted as a curve over time, wherein a physical cleaning is started when a magnitude of a gradient of the curve has reached or passed a curve maximum.
18. A control method according to claim 1, wherein a crossflow in the filter system is set on a basis of a current permeate flow.
19. A control method according to claim 18, wherein the crossflow is increased when the permeate flow is smaller than a predefined limit value, and the crossflow is reduced when the permeate flow lies above a predefined limit value.
20. A control method according to claim 18, wherein the relative energy consumption for a running production cycle is simulated for different possible crossflows and that crossflow, at which the simulated curve for the relative energy consumption reaches the lowest minimum, is selected.
21. A control method according to claim 1, wherein a produced net permeate volume per unit of time is compared to a target volume, wherein a production quantity is reduced on exceeding the target volume and is increased on falling short of the target volume.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
[0036] In the drawings:
[0037] FIG. 1a is a schematic view of a first example of a filter system;
[0038] FIG. 1b is a schematic view of a second example of a filter system amid the use of a crossflow;
[0039] FIG. 2 is a diagram which shows the differential pressure across the filter system plotted over time;
[0040] FIG. 3 is a diagram of the relative energy consumption of a filtration cycle plotted over time;
[0041] FIG. 4 is a flow diagram which shows the evaluation of the duration of a filtration cycle;
[0042] FIG. 5 is a diagram of the differential pressure across the filter system plotted against the number of filtration cycles;
[0043] FIG. 6 is a flow diagram which shows the adaptation of a limit gradient for the relative energy consumption;
[0044] FIG. 7 is a flow diagram which shows the adaptation of a crossflow;
[0045] FIG. 8 is a flow diagram which represents the evaluation of the length of a cleaning procedure;
[0046] FIG. 9 is a flow diagram for an alternative method for determining the points in time for a physical cleaning and a cleaning-in-place;
[0047] FIG. 10 is a detail flow diagram of the procedure in step X in FIG. 9;
[0048] FIG. 11 is a detail flow diagram of the procedure in step XI in FIG. 9;
[0049] FIG. 12 is a diagram of two possible curves for the global relative energy consumption; and
[0050] FIG. 13 is a diagram of the detail XIII from the curves in FIG. 12.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0051] Referring to the drawings, a first filter system, in which the control method according to the invention can be applied, is shown schematically in FIG. 1a. The filter system as a central element comprises a filter element 2. This for example can be a membrane or an arrangement of several membranes. However, another suitable filter element or other suitable filter elements can be applied. Medium to be filtered, for example contaminated water is fed to the filter element 2 from a feed 4, for example from a well, via a feed pump 6. The filtered medium or water at the exit side of the filter element 2 flows into a collection container 8. The arising concentrate 10 which does not pass the filter element 2, is collected or led away in another suitable manner. Moreover, a backwashing pump 12, via which filtered medium, i.e. permeate can be pumped out of the collection container 8 back to the filter element 2 and through this opposite to the filtration direction, in order to backflush the filter element 2, is arranged on the exit side of the filter element 2 in the flow path to the collection container 8. The medium which thereby exits at the entry side of the filter element 2 is discharged with the concentrate 10. Moreover, a flow sensor 14 is arranged in the arrangement in the flow path from the filter element 2 to the collection container 8 and detects the permeate flow rate, or in the opposite direction the backflush flow rate. Moreover, a first pressure sensor 16 is arranged at the entry side of the filter element 2 and a second pressure sensor 18 at the exit side of the filter element, via which sensors the differential pressure P across the filter element 2, i.e. the transfilter or transmembrane pressure TMP can be determined.
[0052] With the design according to FIG. 1b, it is the case of a filter system with a crossflow. A crossflow pump 20 is arranged in this filter system additionally to the design described by way of FIG. 1a, and this crossflow pump circulates the medium to be filtered, via the entry side of the filter element 2, i.e. a crossflow over the filter element is produced, which has the effect that the filter element becomes contaminated more slowly.
[0053] FIG. 2 schematically shows the operation of the filter element 2, wherein the pressure or the pressure difference TMP across the filter element 2 is plotted against time T. It can be recognized that many filtration cycles 22 follow one another, wherein each filtration cycle 22 initially consists of a physical cleaning 24 and a subsequent production phase or a subsequent production cycle 23, in which the actual filtration procedure is carried out. Again a physical cleaning 24 is effected subsequently to the production cycle 23, whereupon the next filtration cycle 22 begins. It can be recognized that the differential pressure TMP across the filter element 2 increases during the production cycle 23 and can be reduced again with the physical cleaning 24 which in particular is a backwashing with the help of the backwashing pump 12 described above, wherein however the initial condition cannot be recreated again. A cleaning-in-place 28 amid the use of cleaning agents or chemicals is effected after a certain number of such filtration cycles 22, by which means the differential pressure across the filter element 2 can be reduced by a greater amount. Subsequently, again several filtration cycles 22 are effected until a renewed cleaning-in-place 28 is necessary. A cleaning-in-place cycle 26 thus includes a cleaning-in-place 28 and a subsequent number of filtration cycles 22 up to the beginning of the cleaning-in-place procedure 28.
[0054] Whereas in known systems, the operation of the filter system is mainly controlled via the changes of the differential pressure TMP across the filter system or the filter element 2, according to the invention, one envisages designing the control method such that the system is operated in an energy-optimized or cost-optimized manner. For this, firstly one envisages controlling the filter system such that a desired net permeate flow Q.sub.N is produced. I.e. the produced permeate quantity minus the permeate quantity which is necessary for the cleaning or backwashing is considered and is held at a desired setpoint. By way of this, one prevents too much permeate being produced, which would lead to unnecessary energy costs and cleaning costs.
[0055] Moreover, according to the invention one envisages detecting the total energy consumption E.sub.G of the filter system and thus the total costs, in a continuous manner during the filtration cycle 22. The total energy consumption E.sub.G as a rule is the electrical energy consumption of the complete system which first and foremost is determined by the energy consumption of the pumps 6 and 12 or, as the case may be, of the crossflow pump 20. The total energy consumption E.sub.P is summed from the energy consumption E.sub.P for the production cycle as well as the energy consumption E.sub.B for the backwashing procedure. Thereby, it is always a filtration cycle 22 beginning with the backwashing procedure which represents a physical cleaning 24 which is considered. The relative energy consumption E.sub.rel is determined from this total energy consumption E.sub.G, by way of the total energy consumption E.sub.G being related to the produced net permeate volume Q.sub.N. Thereby, the net permeate volume Q.sub.N is the produced permeate volume Q.sub.P minus the backwashing volume Q.sub.B which is required for the backwashing procedure.
[0056] The filtration thereby takes its course as is represented in FIG. 4. Beginning with step S1, firstly an estimated backwashing volume Q.sub.B1 and an estimated backwashing energy consumption E.sub.B1 is ascertained in step S2, firstly for the first filtration cycle 22 at the beginning of which no physical cleaning 24 takes place. Then on the basis of these values, the production energy consumption E.sub.P is detected and summed in a continuous manner during the production cycle 23, in step S3. The produced permeate volume Q.sub.P is simultaneously continuously detected and summed. Thus the current relative energy consumption E.sub.rel is continuously determined on the basis of the previously ascertained values for the backwashing energy consumption E.sub.B (for the first cycle E.sub.B1) and the backwashing volume Q.sub.B (for the first cycle Q.sub.B1). FIG. 3 shows the curve of the relative energy consumption E.sub.rel, plotted against time T. One can recognize that this firstly at the beginning is negative and then tends to infinity, since no produced permeate volume Q.sub.P, but only consumed backwashing volume Q.sub.B is present at the beginning. As soon as so much permeate volume Q.sub.P has been produced, that the consumed backwashing volume Q.sub.B is filled up again, the relative energy consumption E.sub.rel decreases over time, until it reaches a minimum 30. The gradient of the curve is equal to zero at the minimum 30.
[0057] The gradient N, i.e. the derivative of the relative energy consumption E.sub.rel is continuously monitored, and examined as to whether the gradient has reached a limit gradient N.sub.STOP, during the production cycle 23, via step S4. The production cycle 23 is continued as long the limit gradient N.sub.STOP has not been reached. If the limit gradient N.sub.STOP is reached, the production cycle 23 is stopped in step S5 and a physical cleaning 24, i.e. a backwashing is carried out. In the step S6, the energy consumption E.sub.B required for the physical cleaning 24 and the required backwashing volume Q.sub.B is detected and then in step S3 forms the basis with the computation of the relative energy consumption E.sub.rel for the next filtration cycle 22. In the example according to FIG. 3, the production cycle 23 is stopped on reaching a limit gradient N.sub.STOP=0, i.e. on reaching the minimum 30. By way of this, one succeeds in the production cycle 23 being completed such that the intervals between the physical cleanings 24 are selected such that a production in the production cycle 22 can be carried out at the minimal energy consumption or cost expense whilst taking into account the necessary physical cleaning 24. However, what is not taken into account thereby is the fact that a cleaning-in-place (CIP) 28 is necessary in certain intervals. This likewise requires energy and causes costs. Inasmuch as this is concerned, it can be desirable to tolerate a higher energy consumption in the individual filtration cycles 22 if the CIP cycle 26 can be extended by way of this and thus the number of necessary cleaning-in-place procedures 28 can be reduced. This is effected in a manner such that the limit gradient N.sub.STOP is adapted whilst taking into account the energy consumption or the costs for a CIP cycle 28, i.e. is displaced by a certain amount with respect to the minimum 30.
[0058] FIG. 6 shows the respective method course or procedure. The filtration is started in step S1, wherein then firstly an estimated energy consumption E.sub.B1 for the first physical cleaning and an estimated backwashing volume Q.sub.B1 for the first physical cleaning are ascertained in step S2 for the first cycle, as described above. Moreover, the counter I for the number of filtration cycles 22 is set to 0 and the limit gradient N.sub.STOP is firstly set to 0. The continuous detection or summing of the relative energy consumption E.sub.rel is effected then in step S3, wherein the backwashing volume Q.sub.B1 and the backwashing consumption E.sub.B1 are taken as a basis for the first filtration cycle 22. Then in each case the backwashing volume Q.sub.B and the backwashing energy consumption E.sub.B of the preceding backwashing procedure or the preceding physical cleaning 24 are taken as a basis for subsequent filtration cycles 22. In step S4, the computation of the gradient N of the curve of the relative energy consumption over time T and the comparison with the limit gradient N.sub.STOP are effected in step S4, as is described by way of FIG. 4. The production cycle 23 is continued for as long as this is not reached. The steps S5 and S6 take their course as described by way of FIG. 4. However yet a step S4B is inserted between the steps S4 and S5, in which step the counter I for the number of filtration cycles 22 is compared to a set point or default value I.sub.CIP. The set point indicates a defined number of cycles (e.g. 10, 20, 30, etc), after which an estimation of the costs or the energy consumption for the complete CIP cycle 26 is begun. As long as this set point I.sub.CIP is not reached, the method continues to run with the steps S5 and S6 as explained by way of FIG. 4. If the set point I.sub.CIP is reached, then in step S7 the relative total energy consumption for the complete CIP cycle is predicted on the basis of the relative energy consumption E.sub.rel summed over several filtration cycles 22. Thereby, the energy consumption or the costs for a preceding cleaning-in-place procedure 28 are taken into account. If a preceding CIP procedure 28 has not yet taken place, then the estimated costs are firstly applied.
[0059] An extrapolation of the curve which is represented in FIG. 5 is carried out, in order to be able to estimate the number of the filtration cycles 22 which are yet to follow until the next CIP 28. In FIG. 5, the differential pressure TMP across the filter element 2 or across the complete filter system is plotted against the number of cycles I. A CIP 28 is started when this differential pressure TMP reaches a predefined limit value. The gradient of the curve over the filtration cycles I is extrapolated up to this limit value, so that the number of filtration cycles yet to be effected until then is estimated. The energy consumption is estimated in a summed manner in step S7 for this. In step S8, it is compared as to whether the thus estimated energy consumption is greater or smaller than the energy consumption which is estimated in the preceding filtration cycle 22. Here, the costs can also be consisted instead of the energy consumption, in which costs for example the costs for the necessary cleaning agents in the cleaning-in-place procedure can also be taken into account. If the currently estimated energy consumption or the currently estimated costs for the complete CIP cycle are smaller than the energy consumption or the costs which were estimated in the preceding filtration cycle 22, the limit gradient N.sub.STOP is reduced by a valve ΔN in step S9. The step S5 described above is then effected subsequently to this. Step S10 then follows if the estimated energy consumption or the estimated costs are not smaller than the energy consumption estimated in the preceding filtration cycle 22 or the costs estimated there, and the limit gradient N.sub.STOP is increased by the predefined valve ΔN. Thus an adaptation of the limit gradient N.sub.STOP is effected and, as the case may be, a shift of the point at which a physical cleaning 24 is started, with respect to the minimum 30 shown in FIG. 3. Thus the point on the curve shown in FIG. 3 can be displaced to the left or to the right. Thus the length of the filtration cycles 22 is optimized in a manner such that as a whole the energy consumption or the costs for the complete CIP cycle 26 can be minimized.
[0060] If the filtration during a filtration cycle is carried out with a crossflow, then according to a particular embodiment of the invention, one envisages likewise optimizing this crossflow. This according to a first method variant can be effected as is shown in FIG. 7. After the start of the control process in step S1, firstly predefined initial values are set in step S2. Thus a change value dQ for the crossflow is set as a predefined value. Moreover, a setpoint for the crossflow Q.sub.crossflowmax is set. A counter I is set to zero and moreover a flow Flux.sub.ref is set as a desired flow for the produced permeate. In the next step S3, the set crossflow Q.sub.crossflow is firstly set to the setpoint Q.sub.crossflowmax which is set in step S2. Since the counter I is equal to zero, no increase or reduction of the value is effected at this point in time. In step S4 in FIG. 7, one then examines as to whether the current permeate flow Flux is smaller than the desired flow Flux.sub.ref. If this is the case, firstly step S5 follows, in which no change of the crossflow Q.sub.crossflow takes place since the counter I continues to be zero. An examination as to whether the permeate flow Flux is smaller than the previously set desired flow Flux.sub.ref is effected afresh in step S6. If this continues to be the case, an increase of the counter I by the value one is effected in step S7 and subsequently an increase in the crossflow Q.sub.crossflowmax by the value i.Math.dQ is thus effected in step S5. Step S6 is effected anew subsequently to this. If the inquiry in step S6 results that the permeate flow is the same or greater than the desired flow Flux.sub.ref, then no further increase of the counter i is effected and the crossflow Q.sub.crossflow remains constant. If the inquiry in step S4 results that the permeate flow Flux is the same or greater than the desired flow Flux.sub.ref, then step S8 follows, in which the counter i is reduced by the value one. Following step S8 is then step S3 again, in which then, with a counter i different to zero, the crossflow Q.sub.crossflowmax is reduced by the change value dQ multiplied by the counter i, since the counter i is then negative. This is effected for so long until the desired permeate flow has been reached. A desired permeate flow Flux.sub.ref can be achieved with a minimal crossflow in this manner.
[0061] A further alternative for optimizing the crossflow is to plot the curve of the local relative energy consumption for the running production cycle over time, as has been described previously. Thus different relative energy consumptions for different crossflows can be simulated in a simulation process, and then one can select that crossflow, at which the simulated curve for the relative energy consumption reaches the lowest minimum. The energy consumption for producing the crossflow is also taken into account in the relative energy consumption via the total energy consumption.
[0062] Finally, according to the invention it is also preferable to optimize the individual backwashing procedures or physical cleaning procedures 24 with regard to their duration. For this, one envisages continuously monitoring a system parameter during the backwashing procedure and completing the backwashing procedure 24 when this system parameter is essentially constant or stable. Thereby, a value which represents the hydraulic resistance of the system is considered as a system parameter. I.e. the physical cleaning 24 is carried out until this hydraulic resistance assumes a stable value or approximates a stable value. This method is schematically represented in FIG. 8. The backwashing procedure or the physical cleaning 24 begins in step S1. A setpoint Q.sub.bw.sub._.sub.set and a maximal value P.sub.bw.sub._.sub.max for the pressure during the backwashing procedure are set in step S2. The flow rate thereby corresponds to the flow rate at the flow sensor 14, through the filter element 2 in the flushing direction. The pressure P.sub.bw corresponds to the pressure between the pressure sensors 16 and 18 during the physical cleaning, i.e. the transmembrane pressure TMP.
[0063] During the physical cleaning, the actual pressure P.sub.bw as well as the actual flow rate Q.sub.bw is continuously detected in step S3. Thereby continuously normalized values P.sub.norm and Q.sub.norm are determined by way of division of the detected pressure P.sub.bw by the maximal pressure P.sub.bwmax as well as by way of the division of the square of the detected flow rate Q.sub.bw by the square of the setpoint for the flow rate Q.sub.bw.sub._.sub.set. A characteristic value C for the hydraulic resistance C=Q.sub.norm.sup.2/P.sub.norm is continuously formed therefrom. The change dC of this characteristic value C according to the formula shown in FIG. 8 in step S4 is determined in step S4, by way of the difference of the characteristic value C and the running average C.sub.avg and this difference being relates to the running average C.sub.avg. The physical cleaning 24 is continued as long as this deviation dC is greater than a predefined value. If the deviation dC reaches or falls short of the limit value, this is an indication that the characteristic value C which represents the hydraulic resistance of the system is stable. The cleaning procedure 24 is stopped in this case in step S6.
[0064] An alternative method procedure for determining the point in time for a physical cleaning 24 or a CIP procedure which is to say a cleaning-in-place 28 is described by way of FIGS. 9-11. Thus both alone can be effected on the basis of the consideration of the global relative energy consumption E.sub.rel or of the curve of the global relative energy consumption E.sub.rel between two CIP procedures. FIG. 12 shows two possible curves of the global relative energy consumption plotted over time. FIG. 13 shows enlarged detail XIII of the curves shown in FIG. 12. The curve of the global relative energy consumption corresponds essentially to the curve for the local relative energy consumption which shown in FIG. 3, but is only plotted over a correspondingly longer time period over several physical cleanings. Furthermore, the curve, for the global relative energy consumption, do not run smooth but as shown in FIG. 12 and FIG. 13 shows deflections or peaks, which indicate the physical cleaning, which leads to a temporary increase of the relative energy consumption. After an effected CIP procedure 28, the method is started in step S1 in FIG. 9. At the beginning, the net permeate volume is firstly Q.sub.N=0. Simultaneously, the total energy consumption E.sub.G up to this point in time corresponds to the energy consumption for the preceding CIP procedure. In step S3, one firstly examines as to whether the produced net permeate volume Q.sub.N is greater than the CIP volume Q.sub.CIP, which means to say the permeate volume which was consumed in the preceding CIP procedure 28. As long as the net permeate volume Q.sub.N is smaller than the previously consumed CIP volume Q.sub.CIP, the evaluation of the point in time for a physical cleaning 24 is effected either as described previously by way of FIG. 4 or represented schematically in FIG. 11. The method which is shown in FIG. 11 corresponds to step S3A or the step S3 in FIG. 4. There, the local relative energy consumption, i.e. for the running production cycle is continuously detected. In step S4A, in contrast to FIG. 4, one does not examine as to whether a limit gradient N.sub.STOP has been reached, but whether the gradient N.sub.N is greater than zero. This mean that here N.sub.stop=0. If the gradient 0 has been reached, then a physical cleaning 24 is carried out in step S5, (also step S5 in FIG. 9). The filtration cycle 22 is continued as long as the limit gradient N.sub.STOP=0 has not been reached.
[0065] In FIG. 9, subsequent to an effected physical cleaning 24 in step S5 is step S6, in which it is examined as to whether the gradient N of the global relative energy consumption has changed its sign (polarity). With the global relative energy consumption E.sub.rel forming the basis here, the total energy consumption since the last CIP procedure 28 including the energy consumption for this CIP procedure 28 and the subsequent physical cleanings as well as the subsequent production cycles forms the basis. Here too, the total net permeate volume QN for this time period since the last CIP procedure 28 forms the basis. In contrast to this, only the local relative energy consumption E.sub.rel for the individual production cycle 23 between two physical cleanings 24 forms the basis of step XI which is described by way of FIG. 4 and FIG. 11.
[0066] Step S3 follows again, as long as the gradient N of the global curve, which is shown in FIG. 12 and FIG. 13 since the last CIP procedure in step S6 has not changed its sign. If there is a change of the sign which means the minimum has been passed or exceeded, then a CIP procedure 28 follows in the next step. The method is started anew in step S1. This minimum corresponds to the minimum 30 in FIG. 3.
[0067] In step S3 in FIG. 9, one differentiated as to whether the net permeate volume Q.sub.N has exceeded the CIP volume Q.sub.CIP, which means that more permeate was produced, than in the preceding CIP procedure. The point in time for a physical cleaning 24, as has been described previously, is determined for as long as this is not the case. As soon as more volume has been produced, step X which is described in FIG. 10, follows step S3. Then the point in time of the physical cleaning 24 can likewise be determined by way of the global curve for the relative energy consumption E.sub.rel since the last CIP cycle, which means to say over several production cycles 23.
[0068] This is effected as described in FIG. 10. After the start of the method procedure, a delay dt of a few seconds is waited in step S3B, subsequently to the step S3 in FIG. 9. On the basis of the current curve of the relative energy consumption, E.sub.rel over time the current gradient of this curve at the current point in time is determined subsequently in step S3C. In step S4B it is then compared as to whether the currently determined gradient is more favorable than the gradient in a preceding computation step, which means before the delay dt. If the gradient is more favorable, then step S3B follows anew. The gradient is more favorable if it is steeper, wherein the gradient is negative since the curve drops monotonously in this region. If the curve course becomes shallower, which means the magnitude of the gradient becomes smaller, the gradient course is less optimal and a physical cleaning 24 in step S5 which corresponds to step S5 in FIG. 9, is subsequent to step S4B. Again step S6 in FIG. 9 is subsequent to this. This means that with this method variant, the point in time for the physical cleaning 24 is determined by way of the global curve for the relative energy consumption over the complete CIP cycle 26. The physical cleaning 24 is started when the gradient or its magnitude becomes smaller, which means the curve becomes shallower, which is to say the reduction of the relative energy consumption slows down. The FIGS. 12 and 13 show two curves I and II for the global relative energy consumption, wherein in the curve II the cycle time between two physical cleanings 24 is shortened compared to curve I, this means the curve decreases steeper is insofar more favorable, since the global relative energy consumption is lower.
[0069] While specific embodiments of the invention have been shown and described in detail to illustrate the application of the principles of the invention, it will be understood that the invention may be embodied otherwise without departing from such principles.
