CH676975A5 - - Google Patents

Description

1

CH 676 975 A5

2

Description

The present invention relates to a method for regulating a wastewater treatment plant according to the preamble of claim 1.

Numerous methods are known for regulating such installations. However, their major drawback is that they provide purely tactical answers to an instant measurement, like a chess player who would only play according to a static assessment of the situation, without taking account of the opposing game. Thus, it is not possible to modify the equation which, in the physico-chemical treatment unit, gives the reagent rates according to the pollution and the objectives set to take into account the functioning of the biological purification. With regard to the latter, moreover, it is not conceivable to stop the supply of a filter before a predictable peak of pollution in order to assure it a longer operation in off-peak periods, and this, taking into account, well obviously, the total filtration capacity. To sum up, we can say that the known type of regulations are purely local and that time is not integrated into them.

Also, the purpose of the present invention is to remedy these drawbacks, and, more specifically, to propose a process of the kind specified above which incorporates the concepts of space and time, which ensures reasoned management and at lower cost. treatment facilities, which, at constant cost, allow the maximum elimination of carbon pollution and which respects safety requirements in the most total way.

According to the invention, these aims, and others which will appear subsequently, are achieved, in a process of the kind specified above, thanks to the fact that the variation of the operating parameters of the installation takes account of both the data which are the estimates of instantaneous pollution at the various levels of the purification treatment, the forecasts of the pollutant load over a period of at least 24 hours and the purification potential of the installation.

It is understood that this constant anticipation of the pollution flow and the purification capacities of the installation, previously unknown, as well as taking into account the risks of accidental pollution, allow the process according to the invention to be largely in within the theoretical limits of all the processes belonging to the prior art.

Regarding the purification potential of the installation, it can be determined from the fouling state of at least one of the filters used during the biological treatment and / or from the quantity of biomass which is present in the floor of activated sludge. This method of determining the purifying potential is remarkably reliable and its use constitutes a significant improvement during the implementation of the present process.

As for the forecasts of the value of the pollutant load which constitute another basic element of the invention, they advantageously extend over a duration which corresponds to a cycle characteristic of the pollution profile in flow and in load, which increases the confidence coefficient attached to the corresponding statistical models. Preferably, these forecasts extend over a period of twice 24 hours, and they are examined by sliding in time over 48 hours, which constitutes a satisfactory compromise between the length of the forecasts and their reliability.

Given the difficulty of the problems posed, where the complexity of the reasoning grows very rapidly with the number of operating rules used, the method according to the invention advantageously uses the techniques of artificial intelligence, that is to say to "expert systems", to act on the operating parameters of the installation according to the data provided in particular by the anticipations specified above and by the estimates of instantaneous pollution at the various levels of the purification treatment.

The vast possibilities offered by data processing and, more precisely, in the case in point, by expert systems, make it possible to integrate into the process according to the present invention various information which it would be very difficult - if not impossible - to operate satisfactorily if he did not use these modern methods of treatment.

The other data which it can thus integrate are, in particular, the evaluation of the operating time of the filters used during biological treatment as a function of mass loads, economic imperatives such as the cost of electrical energy at various times of the day, the evaluation of the mass load of an activated sludge reactor as a function of the pollution already received or to be received, and the previous results which allow the system to improve its performance over time.

Another important aspect of the process according to the invention, as has been said above, lies in the estimates of instantaneous pollution at the various levels of the purification treatment. From this perspective, it is clear that the choice of sensors used and their location in the treatment installation are fundamental elements.

Thus, it is advantageous to estimate the pollution of the water to be treated or during treatment from at least one of the following values: suspended matter (MES), biological oxygen demand (BOD), chemical oxygen demand (COD), nitrogen content and phosphorus content.

These values and their method of determination are well known to specialists in the field, but, according to the present invention, their estimation is preferably carried out on the basis of optical density measurements. Thus, one can use, upstream of the physico-chemical treatment, a pollution estimation unit constituted by a lamellar settling tank, a visible UV absorption spectrophotometer with continuous operation on waste water, and a microprocessor which processes the information received according to a pre-established model to supply it, for example, to the expert system mentioned above.

Likewise, an ap5 can advantageously be used, upstream and / or downstream of the biological treatment.

10

15

20

25

30

35

40

45

50

55

60

65

2

3

CH 676 975 A5

4

similar pollution estimation or "pollutometer" including a UV / visibie absorption spectrophotometer with continuous operation. Such pol-lutometers, as well as the pollution estimation unit described above, are preferably associated with turbidimeters and / or laser granulometers in order to refine the measurements of water pollution being processed.

Advantageously, the purification installation which implements the method of the invention comprises a toxic detector which is mounted upstream of the physico-chemical treatment and the data of which are integrated by the system, in order to be able to carry out a correction by chemical route and / or a short-circuiting of the biological treatment and / or a diversion of the toxic flow towards an intermediate storage tank.

According to a preferred embodiment of the present invention, the parameters which make it possible to act permanently on the operation of the installation also include the stirring energy used during the rapid mixing or “flash-mixing” step. »Of physico-chemical treatment.

The description which follows, and which has no limiting character, will make it possible to understand clearly how the present invention can be put into practice. It should be read against the accompanying drawings, among which:

FIG. 1 is a diagram which represents, in an extremely schematic manner, a purification installation used within the framework of the present invention, as well as some of its organs and control circuits: and:

- Figure 2 is a block diagram which shows the overall structure of the computer system based on the techniques of artificial intelligence which is used for the operation of the installation of Figure 1.

As can be seen in FIG. 1, the raw water to be treated arrives at 1 in the purification installation and it begins by passing through an apparatus for determining the incoming pollution which is generally designated by the reference 10 and which will be described later in more detail. At its exit from this device 10, the water is introduced at 16 into a physico-chemical treatment installation which is designated as a whole by the reference 20 and which can be short-circuited, at least partially, by a pipe 21, called “bypass”, the commissioning of the latter and the adjustment of the water flow which passes through it being ensured by a set of valves of which only one has been shown at 22 by way of very schematic illustration. The position and characteristics of such organs can be easily determined by specialists in the field, and it is therefore not necessary to describe or represent them here as a whole.

In its illustrated embodiment, the physicochemical treatment installation 20 comprises a tank 23 into which 24 are metered quantities of coagulation agents for the colloidal particles which are mixed under rapid stirring or “flash-mixing”. »By means of blades 25. This tank 23 is followed by flocculators equipped with slow agitators 26 and by lamellar decanters with compartments 27, the outlet of the sludge taking place at 28 at the base of the latter. These flocculators and these decanters can also be grouped in a single lamellar settling tank when following the teachings of French patent n ° 85.18 301 of December 11, 1985.

The water coming at 29 from the physico-chemical treatment unit 20 and / or from its by-pass 21 can then be subjected to a biological purification in an installation which is generally designated by the reference 30, which is constituted by a number of tanks 30a, 30b, etc. mounted in parallel, and which, like the physico-chemical treatment installation 20, can be short-circuited, at least partially, by a by-pass pipe 31 provided with valves of which only one has been shown in 32.

Each of the tanks 30a, 30b, etc. can be put into service by opening a valve 33a, 33b, etc., mounted on its inlet pipe. These tanks are advantageously of the type described in French patents No. 76.21 246 of July 12, 1976 and No. 78.30 282 of October 25, 1978, and they then comprise a bed of biologically active carbon which is supported by a perforated plate, into which penetrates , at a suitably chosen height, a plurality of air supply tubes 34a, 34b, etc., respectively, and which is traversed, from top to bottom and at a predetermined speed, by the water it s is to purify. Pipes 35a, 35b, etc. interconnect, respectively, the inlet and outlet of the tanks 30a, 30b, etc., and they each comprise a pump and valves not shown, the reason for which will now be explained.

The coal bed contained in the tanks 30a, 30b, etc. is submerged and fixed. It operates under aerobic conditions, and, starting from a fresh load of coal, it is made biologically active during a preliminary phase lasting a few hours during which the water to be purified is circulated to through the bed, for example in a closed circuit via the pipes 35a, 35b, etc., with simultaneous injection of air, until the effluent is practically clear, that is to say free of colloidal turbidity. Finally, this bed must be cleaned after about a week of operation, when the charge through it reaches a predetermined value, by washing in a fluid bed using an ascending stream of purified water, the sludge being discharged through an overflow in a flow channel. At the outlet 36 of this installation 30, purified water is thus obtained.

All these operations are carried out automatically thanks to commands which emanate from a central computer unit 40 and which act on the corresponding mechanisms. Among these automations, we have simply indicated, very schematically and in phantom, a control 41 acting on the valves 32 which make it possible to more or less short-circuit the biological purification installation 30, individual controls 42 opening or closing the valves 33a, 33b, etc. feed

5

10

15

20

25

30

35

40

45

50

55

60

65

3

5

CH 676 975 A5

6

in water to be purified from corresponding tanks 30a, 30b, * etc., these valves determining the number of treatment stages, additional individual controls 43 governing the flow of air injected into your various tanks, and, for each of these, individual controls 44 making it possible to circulate the water during treatment, in one direction or the other, through the pipes 35a, 35b, etc. As indicated above, these pipes allow a downward circulation of water in the tanks 30a, 30b, etc., for example in a closed circuit, during the preliminary phase intended to make the carbon which they contain biologically active. , and an upward circulation of water during the washing or cleaning phase of the activated carbon.

Other commands act on the operation of the physico-chemical treatment installation 20, such as for example a control 45 acting on the valves 22 which are responsible for short-circuiting this installation more or less, a control 46 regulating the addition of reagent at 24 into the rapid mixing tank 23, and a control 47 modifying the stirring energy of the blades 25 of this Guve,

On the other hand, whether for the physico-chemical treatment installation 20 or for that of biological purification 30, other control members and their means of implementation have neither been described nor represented, in order to avoid any unnecessary overload of the figure and of this description. This is the case with the other valves, pumps and pipes necessary for the operation of the treatment plant which has just been described, because their mounting, moreover within the reach of those skilled in the art, has been widely described. in the aforementioned French patents.

It is now necessary to describe the sensors which are advantageously used for the implementation of the method of the invention, since their nature and their position are extremely important for ensuring the intelligent management of the purification means and of the regulation instruments of the corresponding installation.

We will start with the sensor 10 which is supplied with raw water at 1 and which is installed upstream of the physico-chemical treatment installation 20. This sensor, called ECOFLOC (registered trademark) is the subject of French patent n ° 85.15 245 of October 15, 1985 and will therefore not be described here in detail. It will suffice to indicate that its design results from the observation that, in this case, it is necessary to eliminate, before measuring the pollution, the materials which can settle spontaneously and which do not involve any consumption of reagent during the subsequent purification in the 'physico-chemical purification installation 20. The ECOFLOC therefore consists of a unit 11 for separating decantable materials operating continuously and miniaturized using the principle of lamellar decantation. At its base, the sludge comes out at 12 and it is followed by a spectrophoto-absorption meter 13 UV / vtsibie which is suitable for continuous operation on waste water and whose information is sent by a line 14 represented by mixed lines at the central computer unit 40.

This ECOFLOC, associated with a simple T1 classic type turbidimeter, allows in particular to know the incoming pollution flow, the quantity of sludge that will be produced, as well as an appreciation of the quality of the latter by the MES / iron hydroxide ratio. . In addition, a toxic detector B is advantageously mounted between the ECO-FLOC 10 and the turbidimeter T1 and it can in particular be used to control the diversion of any toxic flows towards an intermediate storage.

An apparatus P2 for determining pollution or "pollutometer" is mounted in the vicinity of the outlet 29 of the installation 20 for physico-chemical treatment. Like ECOFLOC 10, it is based on the principle of measuring the absorption of ultraviolet radiation combined with a measurement of absorption in the visible, without however including any decantation, and it is associated with a turbidimeter T2 which refines the measurements.

Likewise, a pollutometer P3 associated with a turbidimeter T3 is mounted at the outlet 36 of the biological purification unit 30. It should also be specified here that the turbidimeters Tt, T2 and T3 could be replaced by laser granulometers of known type.

As shown very schematically at 47 in FIG. 1, the information provided by the sensors B, T1 to T3, P2 and P3 are brought to the central computer unit 40, and it should be emphasized here that the configuration of these measuring devices, as just described, is that which allows both the best regulation of the processing parameters at a given time, and provides the most relevant data for the application of the inference rules of the central unit's expert system 40.

Finally, many other data can be introduced into this central unit 40, such as temperatures, fluid flow rates, and, more importantly, knowledge bases. Except for the latter, the case of which will be discussed at length below, they will not be described here in detail, and their introduction into the central unit 40 has been simply shown at 48.

To arrive now at the expert system which is used within the framework of the present invention and which is implemented in the central unit 40, it will be recalled first of all that such a system, which is also called “based on knowledge ”and which calls on the very recent techniques of artificial intelligence, is a computer tool whose function is to simulate the reasoning of a human expert and which is distinguished from a traditional software by the separation of the parts“ reasoning ” and "knowledge". Compared to a classical program which can only solve problems that are always rigorously posed in the same way, a program written in "expert system" does not constitute a sequence of instructions, but a collection - not sequential, but nevertheless structured - of assertions used by means of inference processes which replace the algorithmic and repetitive processes of usual data processing. The data are thus supplemented by knowledge5

10

15

20

25

30

35

40

45

50

55

60

65

4

7

CH 676 975 A5

8

We also know how to represent, manipulate and use to lead to reasoned decisions.

Referring to FIG. 2, such a knowledge-based system, S, is provided with an inference engine M which is a program constructing reasonings from a fact base BF. In this area, the definition of the word "inference" is that given by any dictionary, namely an intellectual operation by which one passes from one truth to another truth, judged as such because of its link with the first.

The inference engine M provides its conclusions to an interface I for interaction with the user which provides a synoptic of the results achieved and which even allows, possibly, to force the parameters of the purification treatment from the knowledge stored in the system S. It is thanks to this interface I that the process is carried out as a whole in P, including the fixing of the intermediate objectives, by action on the commands 41 to 47 and 14 of FIG. 1 and on all the other commands which were simply mentioned above.

At the level of interface I, the whole process is displayed graphically on a screen and the positioning of a cursor on such or such point of the system makes it possible to call the instantaneous values of the flows, as well as their history or their forecast. . The positioning of this cursor on an organ such as a valve or a motor allows, for its part, to carry out a change of state by remote control. In addition, the operator can visualize at any time the "reasoning" of the system, or at least its logical approach, this "user-friendly" aspect of dialogue with the machine is not the least interest of the process which provides the user a kind of “intelligent” screen.

The inference engine M is fed by facts and by knowledge which respectively come from a fact base BF and from a knowledge base BC and which are respectively examined by a unit F1 for examining the facts and by a C1 rule review unit. In turn, he can provide the BF fact base with inferred facts that are developed by an FD unit. The BF fact base is materialized by a working memory and it stores in particular the targeted objectives, for example in terms of residual pollution and usable forecasts. These relate in particular to the consumption of water and its expected pollution, to the data supplied by the sensors, and to the inferred facts provided, as has just been said, by the inference engine M itself. .

As for the BC knowledge base, it is responsible for all the "meta-knowledge" that can be brought to it by experts in the field, particularly with regard to the mode of operation of the physico-chemical and biological treatment units, the evacuation of sludge, the interactions between the various phases of the purification treatment, and many more.

This knowledge is not only brought in the form of rules, but also in that of frames or "frames" which are sometimes defined as the concrete computer representation of a context and which makes it possible to approach the hierarchy of objects simply. Each framework is accompanied by a set of waiting values, by default, and the rules are hierarchical so as to be able to distinguish, in particular, simple rules from “meta-rules” which make it possible to impose a strategy on the system.

Finally, as in all expert systems, this information can be provided by specialists who are consulted in the form of production rules, that is to say expressions of the form "if conditions, then conclusions" , which allows the specialist to express his expertise in a simple and modular way and also facilitates the updating of the BC knowledge base.

The system S is also supplied by models which have been grouped together in FIG. 2 in a MOD block and which integrate, for example, the forecast of the pollution flow and the various operating parameters such as the reagent rates, the cycle time. biological treatment tanks and many more. The system S is also supplied by a database BD which stores the entries ENT constituted by the values supplied by the various sensors of the installation. These values can be those of the weight of suspended matter (MES), of the chemical oxygen demand (COD), of the biological oxygen demand (BOD), of the flow rates, of the temperatures, of the optical densities giving inter alia the contents in nitrogen and phosphorus, and others. The BD database can also store treatment parameters and the malfunctions that may be encountered.

It is very clear that the system S can easily integrate the essential data of the regulation process according to the present invention, namely the data which are constituted, in a perfectly original manner, by the estimates of instantaneous pollution at the various levels of the treatment of purification, forecasts of the pollutant load over a period of at least 24 hours, and the purification potential of the installation. This data is supplied to it, in particular, by the database BD and by the MOD models which are both supplied by the inputs ENT. They constitute a part of the fact base BF, and, thanks to the knowledge of the knowledge base BC, they allow an optimal control of the process P by comparison and correction.

Of course, only the possibilities offered by expert systems allow such optimization, given the enormous volume of data and “meta-knowledge” to be implemented. On the pollution flow forecasts, it should also be recalled that it has been shown that it is possible to predict the pollution flow during a day by taking into account the pollution profile of the previous day, the month and the day of the week you are in, and certain elements of the climate in the more or less long term. Modeling of this type can be carried out on all sites provided there is one year of continuous monitoring, and it is stored in the MOD unit.

5

10

15

20

25

30

35

40

45

50

55

60

65

5

9

CH 676 975 A5

10

Be that as it may, such modeling makes it possible to carry out quite easily an important arrangement of the method according to the present invention, namely that the forecasts of the value of the pollutant load must extend over a duration which corresponds to a characteristic cycle. of the pollution profile in terms of flow and load, these forecasts preferably extending over a period of twice 24 hours and being examined by sliding over time over the 48 hours following the analysis.

However, other parameters which can also be taken into account for feeding the knowledge-based system S are for example:

- The stirring energy of the blades 25 (see FIG. 1) of the rapid mixing installation 23;

The operating time of the filters used in the biological purification installation;

- Economic imperatives, and, in particular, the cost of electrical energy, given the existence of so-called “off-peak” hours when electricity is cheaper;

- The data provided by the toxic detector B which plays a special role vis-à-vis the other sensors of the installation, because it controls operations of another nature.

Finally, the interactions between the knowledge-based system, S, the BD database and MOD models allow the system to integrate previous results to improve its performance. Thus, from sensors P2 and T2 (Figure 1), for example, we can continuously enrich the database BD with new information thus establishing the relationship:

Treatment parameters = f (characteristics of the raw water // objectives).

This BD database, which is initialized by data from the knowledge that manufacturers have of the performance of their equipment, therefore becomes more reliable as it operates and thus adapts to the specific qualities of the raw water on a site. given, to the performance of the lamellar decanter 27 (FIG. 1) and to the equipment 24 for injecting reagents.

As has been said above, an important point of the process according to the invention relates to the operation of the filters used in the biological purification installation, and the following example, taken from the management of the washing of six filters, will demonstrate the complexity of such a problem.

Assuming a period which allows washing on average to be carried out in 36-hour increments, a first simple rule leads to proposing a period of 24 hours on four filters and 48 hours on the other two.

At time t = 10 hours, the history of the admitted mass and hydraulic loads, as well as the forecasts for future loads, make it possible to establish the probable durations before washing the following table;

Filter # 1 2 3 4 5 6

Duration (h) 2 7 40 20 13 36

In addition to this configuration, the rules of the filtration unit indicate that the peak expected at noon can be absorbed on five filters, so that it is possible to stop feeding on one filter, and that filter no. 2 must be stopped until 4 p.m. in order to be washed during off-peak periods. In the event of filter # 1 stopping, the following new forecast is established:

Filter # 1 2 3 4 5 6

Duration (h) X 5 35 17 11 32

We deduce from these values that the filter n ° 2 must be washed at 3 p.m. or stopped; but, by applying the rules of the filtration unit, it cannot be stopped without risking exceeding the maximum load during the peak. A new overall combination is then sought by fixing an increase in yield on the physico-chemical treatment installation and by evaluating the consequences on the quality of the treatment and on the costs. A new cycle of reasoning is then started, and we see that the complexity of this problem - which is however not the most complicated to solve - justifies that, according to the invention, we call upon an expert system.

In fact, this system makes it possible to determine the purification potential of the biological purification installation from the fouling state of the filters used, and to integrate the evaluation of the operating time of the latter as a function of the loads. mass and hydraulics that are applied.

More specifically, the biological filtration stage receives the residual pollution flow after treatment by the physico-chemical treatment installation. This flow is also characterized by a flow rate, a MES concentration and a BOD concentration which are measured by the sensors P2 and T2 (Figure 1) located at the exit of the previous stage.

The operating parameters of each of the filters to reach a given target in MES and in BOD output are then the instantaneous speed of the water on the filter, which results inter alia from the number of filters in service, and the air flow to be injected which results from a preGÎs calculation from the pollution load.

In the system of the invention, each filter keeps track of its activity since the previous wash in the form of a degree of clogging and a potential for cycle time before the next wash. It is clear that these factors depend simultaneously on the MES retained on the filter and on the BOD already eliminated by the biological activity. Experience also makes it possible to integrate, for the knowledge of the cycle time, the concept of pollution profile received or to be received which was mentioned above.

This corresponds to the intuitive notion that the same pollution load corresponding to a given volume of water to be treated does not have the same impact, both on the duration of the filtration cycle as on the qualities of the treated water, depending on whether it is distributed over a fixed debit or whether it is processed

5

10

15

20

25

30

35

40

45

50

55

60

65

6

11

CH 676 975 A5

12

with both hydraulic and concentration tips.

In the same way as for the physico-chemical stage, one can, from the sensors located on the treated water, permanently compare the results obtained and the expected results and enrich the BD database, subject to certain rules. relatively complex, however, so as to obtain the relationships:

Treatment parameters = f (treated water, filtered water objectives).

At each instant, each filter is thus characterized by its operating parameters and by a potential cycle time with regard to the pollution profiles to be received.

It should also be noted that the purification potential of the installation can also be determined from the quantity of biomass present in the stage of the activated sludge of the biological purification installation, and that the expert system used in the context of the present invention can integrate the evaluation of the mass load of the activated sludge reactor as a function of the pollution already received or to be received.

Finally, it will be recalled that the inference engine of the expert system can use two types of reasoning commonly called "front chaining" and "back chaining". With the first, reasoning is done using the rules from known facts until the goal has been deduced. In back chaining, on the contrary, the reasoning takes place from the goal while trying to bring it back to equivalent sub-goals that we then seek to verify.

The expert system implemented within the framework of the present invention can use these two types of reasoning: forward chaining when, starting from the quality of raw water, it defines the operating parameters of the physical treatment installation -chemical, and back chaining when, from the treatment objective, the state of the filters and the pollution profile to be received, it defines the characteristics of the decanted water that it must receive.

To conclude, we see that the regulation process according to the invention makes it possible to constantly react to the reality of the effluents arriving on a treatment plant, and this, depending on the state of the latter and by integrating its operation. in time. This permanent adaptation in particular opens the way to several types of improvements in addition to those already described. Thus, the quality of the treatment is improved by permanently using each item of equipment provided to the best of its ability, on the one hand, and, on the other hand, we manage to not always use all the equipment provided.

For a given treatment capacity, a higher overall reliability is obtained and the installation is given the capacity to mobilize either an additional physicochemical treatment or a certain number of filters, until then stopped. , to deal with an unforeseen instantaneous pollution load compared to the standard pollution profiles.

Claims (14)

Claims 1. Method for the regulation of a wastewater treatment plant using a physico-chemical treatment and a biological treatment which can operate in parallel or in series, and in which one acts permanently on the residual pollution flow by making vary at least one of the parameters which are the quantity of flocculation reagent added to the water to be treated, the distribution of the water flows between the biological treatment and the physico-chemical treatment, as well as the air flow injected during the biological treatment, and the number of stages of treatment, characterized in that the variation of said parameters takes into account both the data that are the estimates of instantaneous pollution at the various levels of the purification treatment, the forecasts of the pollution load over a period of at least 24 hours and the purification potential of the installation. 2. Method according to claim 1, characterized in that said purifying potential is determined from the fouling state of at least one of the filters used, during biological treatment and / or from the amount of biomass present in the stage of activated sludge. 3. Method according to any one of claims 1 and 2, characterized in that said forecasts of the value of the pollutant load extend over a duration which corresponds to a cycle characteristic of the pollution profile in flow and in load. 4. Method according to claim 3, characterized in that said forecasts of the value of the pollutant load extend over a period of twice 24 hours, and that they are examined by sliding in time over 48 hours. 5. Method according to any one of claims 1 to 4, characterized in that one uses the techniques of artificial intelligence to act on said parameters from said data. 6. Method according to any one of claims 1 to 5, characterized in that it integrates the evaluation of the operating time of the filters used during said biological treatment as a function of the mass and hydraulic loads applied. 7. Method according to any one of claims 1 to 6, characterized in that it integrates the evaluation of the mass load of an activated sludge reactor according to the pollution already received or to be received. 8. Method according to any one of claims 1 to 7, characterized in that the said parameters also include the stirring energy which is used during the physico-chemical treatment. 9. Method according to any one of claims 1 to 8, characterized in that the pollution of the water to be treated or during treatment is estimated from at least one of the values of the suspended matter (MES ), biological oxygen demand (BOD), chemical oxygen demand (COD), and nitrogen and phosphorus contents. 10. Method according to claim 9, characterized in that the estimation of the MES, BOD values 5 10 15 20 25 30 35 40 45 50 55 60 65 7 13 CH 676 975 A5 and COD, and nitrogen and phosphorus contents is carried out from optical density measurements. 11. Method according to any one of claims 1 to 10, characterized in that one uses, upstream of said physico-chemical treatment, a unit for estimating the pollution constituted by a lamellar descaler, a spectrophotometer UV / visible absorption system with continuous operation on waste water, and a microprocessor which processes Your information received according to a pre-established model. 12. Method according to any one of claims 1 to 11, characterized in that one uses, upstream of said physicochemical treatment, a toxic detector whose data are integrated for correction by chemical means and / or for short-circuiting the biological treatment and / or for diverting the toxic flow to an intermediate storage tank. 13. Method according to any one of claims 1 to 12, characterized in that one uses, upstream and / or downstream of said biological treatment, a pollution estimation apparatus comprising a UV absorption spectrophotometer / visible with continuous operation. 14. Method according to any one of claims 1 to 13, characterized in that one uses, upstream and / or downstream of said physico-chemical treatment and / or said biological treatment, a turbidimeter and / or a particle size analyzer laser to estimate the pollution of the water to be treated. 5 10 15 20 25 30 35 40 45 50 55 60 65 8