GB2208856A - Regulating installations for treating waste water - Google Patents

Description

k.

90/3181/01 2."/-'j' 8 b o 1 PROCESS FOR REGULATING INSTALLATIONS FOR PURIEYING WASTE WATER.

DESCRIPTION

The present invention relates to a process for regulating installations for purifying waste waters which employ a physicochemical treatment and a biological treatment which can operate in parallel or in series.

In these installations, in a conventional manner, the residual flow of pollution is continuously influenced by varying at least one of te parameters comprising the quantity of flocculant added to the water to be treatedr the distribution of the flow of water between the biological treatment and the physicochemicaltreatment, and the flow rate of air injected during the biological treatment, and, where applicable, the number of stages of treatment.

Numerous methods are known for regulating installations of this kind. Neverthelessr their major disadvantage is that they provide purely tactical responses to an instantaneous measurement, like a chess player who plays according to a static evaluation of the situation without taking account of his opponent's game. Thus it is not possible to modify the equation which, in the physicochemical treatment unity gives the proportions of reagents as a function of the pollution and of the objectives determined so as to take account of the operation of the biological purification unit. As regards the latter, furthermorer it is not possible to envisage stopping the supply to a filter before a foreseeable level of pollution to ensure that it has a longer operational time in a slack periody taking accountr of course, of the total filtration capacity. In summary, it can be stated that the regulation systems

90/3181/01 1 i 2 of the known type are purely spatial and that time element has not been integrated into them.

Thus. the object of the present invention is to remedy these disadvantages andr more particularly, to propose a method of the type specified above which includes thp notions of space and time, ensures rational and low cost management of the treatment installations whichr at a constant costy permits maximum eliminat.ion of pollution containing carbon and complies fully with the imperatives of safety.

According to the present invention a process for regulating installations for purifying waste water. these installations being of the type which employs a physicochemical treatment and a biological treatment which can operate in parallel or in series. and of the type in which the residual flow of pollution is continuously influenced by varying at least one of the parameters comprising the quantity of flocculant added to the water to be treated. the distribution of the flow of water between the biological treatment and the physicochemical treatment, and the flow rate of air injected during the biological treatment. and. where applicable. the number of stages of treatment. is characterised in that the variation of the parameters simultaneously takes account of the data comprising the assessments of the instantaneous pollution at the various levels of the purification treatmentr the forecasts of the pollution load over a period of at least 24 hours and the purification potential of the installation.

It will be understood that this constant anticipation of the pollution flow and of the purification capacity of the installationj which anticipation was unknown hitherto, and the taking into account of the risks of accidental pollutionj permit the process according to the invention to remain broadly within the theoretical limits to which 1 the 90/3181/01 1 3 all the processes belonging to the prior art are subject.

As regards the purification potential of the installation, it can be determined on the basis of the soiling of at least one of the filters employed during the biological treatment and/or from the quantity of biomass which is present in the activated sludge stage. This method of determining the purification potential is remarkably reliable and its use constitutes an important improvement when the present process is implemented As regards the forecasts of the value of the pollution loadr which form another fundamental element of the invention, these advantageously extend over a period which corresponds to a cycle characteristic of the pollution profile in terms of flow rate and content, and this increases the level of confidence attached to the corresponding statistical models. These forecasts preferably extend over a period of two times 24 hours, and they are examined after a time lapse of 48 hours, this being a satisfactory compromise between the length of the forecasts and their reliability.

Given the difficulty of the problems posed, in which the complexity of the logic increases very rapidly with the number of operating rules employed, the process according to the invention advantageously has recourse to the techniques of artificial intelligence, i.e. to "expert systems", to influence the operating parameters of the installation as a function of the data provided in particular by the forecasts specified above and by the assessments of the instantaneous pollution at the various levels of the purification treatment.

The enormous possibilities offered by data processing and, more particularly, in the case in point, by expert systemst make it possible to 90/3181/01 1 4 integrate a variety of information into the process according to the present invention which would be very difficult - if not impossible - for it to make use of satisfactorily if it did not employ these modern processing methods.

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

Another important aspect of the process according to the invention, as stated abover is the assessment of the instantaneous pollution at various levels of the purification treatment. From this viewpoint it is clear that the choice of sensors used and their positioning in the treatment installation are fundamental elements.

Thus it is advantageous to assess the pollution of the water to be treated or in the process of treatment on the basis of at least one of the following values: the matter in suspension (MIS)r the biological oxygen demand (BOD), the chemical oxygen demand (COD)p the nitrogen content and the phosphorus content.

These values and the way in which they are determined are well known to specialists in this field, but, according to the present invention, they are preferably assessed on the basis of optical density measurements. Thus, upstream of the physicochemical treatment unit for assessing the lamellar settler#, it is possible to use a pollution comprising a a Uv/visible absorption Q 90/3181/01 spectrophotometer operating continuously on waste water and a microprocessor which processes the information received in accordance with pre-established model and feeds them, for example, to the expert system discussed above.

Likewise, it is possible to use with and/or downstream of the advantage, upstream biological treatment# a device for assessing the pollution or "pollutometer" comprising a continuously operating UV/visible absorption spectrophotometer. Such pollutometers, as well as the unit for assessing the pollution described abover are preferably associated with turbidimeters and/or laser instruments for determining particle size, for the purpose of refining the measurements of the pollution of the water undergoing treatment.

The purification installation which implements the process of the invention advantageously comprises a toxic substance detector which is mounted upstream of the physicochemical treatment, the data from which are incorporated by the system in order to allow a correction by chemical means and/or short circuiting of the biological treatment and/or diversion of the toxic flow towards an intermediate storage reservoir.

According to a preferred embodiment of the present invention, the parameters which permit the operation of the installation to be continuously influenced further comprise the stirring energy employed during the 'flash mixing" stage of the physicochemical treatment.

In the drawings:

Figure 1 is a diagram which represents, in a highly schematic form, a purification installation employed within the framework of the present invention and certain of its components and control circuits; and Figure 2 is a block diagram which shows the 90/3181/01 4 6 overall structure of the data processing system, which is based on the techniques of artificial intelligence and is used to operate the installation in Figure 1.

As can be seen in Figure ly the raw water to be treated enters the purification installation at 1 and begins by passing through an apparatus for determining the pollution on entry#, which is designated overall by the reference numeral 10 and will be described in greater detail below. At the outlet of this apparatus 1OF the.water is fed in at 16 to a physicochemical treatment installation which is designated overall by the reference numeral 20 and can be short circuited. at least partiallyr by a line 21r called a wbypass line", the bringing into operation of the latter and the regulation of the flow of water which passes through it being ensured by a set of valves only one of which has been represented at 22 by way of highly schematic illustration. The position and characteristics of such components can easily be determined by specialists in this field and it is therefore unnecessary to describe them or depict them here in their entirety.

In the embodiment illustrated, the physicochemical treatment installation 20 comprises a vessel 23 into which. at 24r metered quantities of agents for coagulating the colloidal particles are introducedr these being mixed with rapid stirring or nflash mixingw by means of blades 25. The vessel 23 is followed by flocculators equipped with slow stirrers 26 and with lamellar settlers with compartments 27y the sludge being removed at 28 at the base of the latter. The flocculators and settlers may be gathered together in a single lamellar settling tank according to the teaching of French Patent 85/18301 of 11 December 1985.

90/3181/01 7 The water coming at 29 from the physicochemical treatment unit 20 and/or from its bypass 21 can then be subjected to biological purification in an installation which is designated overall by the reference numeral 30 and is made up of a certain number of tanks 30a, 30b, etc., arranged in parallel, and which, like the physicochemical treatment installation 20, can be short circuited, at least partially, by a bypass line 31 provided with valves, only one of which has been represented at 32.

Each of these tanks 30a, 30b, etc., can be brought into use by opening a valve 33a, 33b, etc.y mounted on its inlet line. These tanks are advantageously of the type described in French Patents 76/21246 of 12 July 1976 and 78/30282 of 25 October 1978 and thus comprise a bed of biologically active carbon which is supported by a perforated plate through which pass, at a suitably chosen height, a plurality of air addition tubes 34a, 34b, etc., respectively, and which is traversed, from top to bottom at a predetermined rate, by the water which is to be purified. Lines 35a, 35b, etc. link the inlet and outlet respectively of tanks 30a, 30b, etc., and each comprises a pump and valves (not shown) the reason for which will be explained now.

The bed of carbon contained in the tanks 30a, 30b, etc. is submerged and fixed. It operates under aerobic conditions and, starting from a fresh batch of carbon, is rendered biologically active in the course of a preliminary phase lasting for several hours during which the water to be purified is circulated through the bed, for example in a closed circuit via lines 35a, 35b, etc., with the simultaneous injection of air, until the effluent is virtually clear, i.e. without colloidal turbidity. Finally, this bed must be cleaned after about a week in operation, when the pressure across it reaches a 90/3181/01 8 predetermined value. by washing in a fluidised bed by means of a rising current of purified water, the sludge being discharged via an overflow into a discharge channel. Purified water is obtained at the outlet 36 of this installation 30.

All these operations are carried out automatically by virtue of commands which are issued by a central processing unit 40 and act on the corresponding mechanisms. Among these automatic functions, the only ones which have been indicatedr and this in a highly schematic manner and with chain dotted linesp are: a command 41 acting on the valves 32 which allow the biological purification installation 30 to be short circuited to a greater or lesser extent; individual commands 42 which open or close valves 33ar 33br etc., for feeding the corresponding tanks 30a, 30b, etc. with water to be purified, these valves determining the number of stages of treatment; additional individual commands 43 which regulate the flow of air injected into the various tanks; and, for each of the latter, individual commands 44 allowing the water undergoing treatment to be circulated through the lines 35a.

35b, etc., in one direction or the other. As indicated above, these lines permit circulation of water from above into the tanks 30a, 30bi etc.r for example in a closed circuit, during the preliminary phase intended to render the carbon which they contain biologically active, and circulation of water from below during the phase in which the active carbon is washed or cleaned.

Other commands act on the functioning of the physicochemical treatment installation 20, for example a command 45 acting on the valves 22 responsible for short circuiting this installation to a greater or lesser extent, a command 46 which regulates the addition of reagent at 24 to the rapid 1 1 90/3181/01 S 9 mixing tank 23, and a command 47 which modifies the stirring energy of the blades 25 in this tank.

other command elements, whether they are for the physicochemical treatment installation 20 or for the biological purification installation 30y and the means for implementing them, have not been described or illustratedy in order to avoid any unnecessary cluttering of the figure and of the present description. There are thus other valves. pumps and lines necessary for the operation of the purification station just described, since their installation. which is in any case within the reach of the person skilled in the art, has been amply described in the above mentioned French patents.

It will now be appropriate to describe the sensors which are advantageously used for implementing the process of the invention since their nature and their position are extremely important to ensure the intelligent management of the means of purification and of the instruments for regulating the corresponding installation.

The sensor 10 is fed with raw water at 1 and is installed upstream of the physicochemical treatment called ECOFLOC of installation 20. This sensor, (registered trademark) is the subject French Patent 85/15245 of 15 October 1985 and will therefore not be described here in detail. It will suffice to point out that its conception results from the observation that in this case it is necessary before measuring the pollution to eliminate the matter which is capable of settling spontaneously and does not entail the consumption of any reagent during the subsequent purification in the physicochemical purification installation 20. The ECOFLOC is thus made up of a continuously functioning unit 11 for separating settlable mattert which unit is miniaturised by virtue of the principle of lamellar 90/3181/01 i settling. At its base, the sludge flows out at 12 and it is followed by a W/visible absorption spectro photometer 13 which is suitable for continuous functioning on waste water and the information from which is transmitted via a line 14 represented by -dots and dashes to the central processing unit 40.

This ECOFLOC, which is associated with a simple turbidimeter Tl of conventional typer permits in particular the incoming flow of pollution and the quantity of sludge produced to be known and permits an assessment of the quality of the latter by means of the MIS/iron hydroxide ratio. Furthermore. a toxic substance detector B is advantageously installed between the ECOFLOC 10 and the turbidimeter Tl and can in particular serve to control the diversion of any toxic flow to an intermediate storage facility.

An apparatus P2 for determining the pollution or wpollutometer" is installed in the vicinity of the outlet 29 of the physicochemical treatment installation 20. Like the ECOFLOC 10r it is based on the principle of a measurement of the absorption of an ultraviolet emission combined with a measurement of absorption in the visible spectrum, albeit without including prior settling, and is associated with a turbidimeter T2 which refines the measurements.

A pollutometer P3 associated with a turbidimeter T3 is likewise installed at the outlet 36 of the biological purification unit 30. The turbidimeters T1, T2 and T3 may be replaced by laser instrumentst for determining particle size, of a known type..

As has been represented in a highly schematic manner at 47 on Figure 1, the information provided by the sensors B7 Tl to T3, P2 and P3 is supplied to the central processing unit 40 and it is appropriate to emphasise here that the configuration of the 90/3181/01 11 1 1 measurement apparatuses, as just described, is the one which permits at the same time the optimum regulation of the treatment parameters at a given instant and provides the most pertinent data for the application of the rules of inference of the expert system of the central unit 40.

Finally, numerous other data items can be introduced into this central unit 40, such as temperatures, fluid flows and, even more important, the data bases of knowledge. Apart from the latter, which will be discussed at length below, these will not be described in detail here, and their introduction into the central unit 40 is thus schematically shown at 48.

To move on now to the expert system which is used within the framework of the present invention and is implemented in the central unit 40, it will be recalled first of all that a system of this kind, which is also called "knowledge-based" and draws on most recent techniques in artificial intelligence, is a data processing tool whose function is to simulate the reasoning of a human expert and is distinguished from a traditional software package by the separation of the "reasoning' and "knowledge" elements. Compared to a traditional program, which can only resolve problems which are set in rigorously identical fashion every time, a program written as an "expert system" does not comprise a sequence of instructions but comprises a non-sequential but nevertheless structured collection of assertions used by means of processes of inference which replace the algorithmic and repetitive processes of conventional data processing operations. The data are thus complemented by knowledge which can with equal ease be represented, manipulated and used to lead to reasoned decisions.

With reference to Figure 2, a knowledge-based the 1 90/3181/01 4 12 system S of this kind is provided with an inference motor M which is a program which constructs reasoned arguments on the basis of a fact base BF. In this fieldr the definition of the word linferencew is the one giveh in any dictionary, namely an intellectual operation by means of which a progression is made from one truth to another truth judged to be such by reason of its connection with the first truth.

The inference motor M supplies its conclusions 10 to a user interface I which delivers a synopsis of the results achieved and, if desired, even permits the parameters of the purification treatment to be overridden on the basis of the knowledge stored in the system S. It is by virtue of this interface I that the conduct of the overall process is effected at P, including the fixing of the intermediate objectives, by acting upon commands 41 to 47 and 14 in Figure 1 and on all the other commands which have been mentioned above.

At the interface I# the overall process is displayed graphically on a screen and the positioning of a cursor on a particular point of the system permits the instantaneous values of the flows, as well as their previous and forecast values, to be called. The positioning of this cursor on a component such as a valve or a motor for its part permits a change of state by remote control to be effected. Moreovery the operator can at any moment display the "reasoning' of the systemr or at least its logical progression, this "friendly" aspect of dialogue with the machine being not the least important of the process which supplies the user with a sort of 'intelligent" synopsis.

The inference motor M is fed with facts and 35 knowledge which have come respectively from a fact base BF and a knowledge base BC and are examined respectively by a fact examination unit F1 and by a Q 90/3181/01 1 13 rule examination unit Cl. In its turn, it can supply the fact base BF with deduced facts, which are elaborated by a unit FD. The fact base BF is realised by a scratch memory and it stores in particular the objectives aimed at, for example in terms of residual pollution and - the usable forecasts. The latter relate in particular to the consumption of water and its predicted pollution, to the data provided by the sensors and to the deduced facts provided, as just mentionedr by the inference motor M itself.

With regard to the knowledge base BCp it is loaded with all the "metaknowledgel with which it can be supplied by the experts in the field concerning in particular the way in which the physicochemical and biological treatment units operate, the evacuation of the sludge, the interactions between the various phases of the purification treatment and many others.

This knowledge is not only supplied in the form of rules but also in the form of "frames", which are sometimes defined as the concrete representation in computing terms of a context and permit a simple approach to the hierarchy of the subjects. Each frame is accompanied by a set of expected default values and the rules are arranged in a hierarchy to enable, in particular, the simple rules to be distinguished from the Ometarules' which permit a strategy to be imposed on the system.

Finallyi as in all expert systemsy this information can be supplied by specialists in the form of production rules, i.e. expressions of the form "if conditions, then conclusions", enabling the specialists to express their expertise in a simple and modular manner and thus facilitates the construction of the knowledge base BC The system S is likewise fed by models which have been gathered together in Figure 2 in a block 90/3181/01 14 MOD and whichr for example, incorporate the forecast of the flow of pollution and the various operating parameters, such as the proportions of reagents, the cycle time of the biological treatment tanks and various others. The system S is likewise fed by a - data 'base DD which stores the inputs ENT made up of the values supplied by the various sensors of the installation. These values may be those of the weight of matter in suspension (MIS)r the chemical oxygen demand (COD)#, the biological oxygen demand (BOD), flow rates, temperaturesi optical' densities giving inter alia the contents of nitrogen and of phosphorus, and various others. The data base BD can likewise store treatment parameters and the malfunctions which may be encountered.

It is quite clear that the system S can easily incorporate the essential data of the regulation process according to the present invention, namely the data which consisty in a completely original manner, of the assessments of the 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. It is provided with these data, in particularr by the data base BD and by the models MOD, both of which are fed by the inputs ENT. They make up part of the fact base BF and, by virtue of the knowledge of the knowledge base BCr they allow the process P to be conducted in an optimum manner by comparison and correction.

Of course, only the possibilities offered by expert systems allow such an optimisation, given the enormous volume of data and of Imetaknowledge' to be employed. on the subject of the forecasts of the flow of pollutionr it will be remembered that it has been demonstrated that it is possible to predict the flow of pollution in the course of a day by taking i 90/3181/01 account of the pollution profile of the previous day. of the month and of the particular day of the week and of certain more or less long term elements of the climate. Modelling of this kind can be carried out at all the sites provided a year is available for continuous monitoring, and it is stored in the MOD unit.

Whatever the case may bef modelling of this kind permits an important provision of the process according to the present invention to be realised with relative ease, namely that the forecasts of the value of the pollution load must extend over a period which corresponds to a characteristic cycle of the pollution profile in terms of flow rate and content, these forecasts preferably extending over a period of two times 24 hours and being examined after a time lapse of 48 hours following the analysis.

This being so, other parameters which may likewise be taken into account for supplying the knowledge-based system S are, for example:

the stirring energy of the blades 25 (see Figure 1) of the rapid mixing installation 23; the operating time of the filters used in the biological purification installation 30; the economic imperatives and, in particular, the cost of electrical energy taking into account the existence of "slack' periods when electricity is less expensive; and the data provided by the toxic substance detector B, which plays a role distinct from the other sensors in the installation by reason of the fact that it controls operations of a different nature.

Finally, the interactions between the knowledge-based system S, the data base BD and the models MOD enable the system to incorporate previous results in order to improve its performance. Thus, 90/3181/01 4 16 using sensors P2 and T2 (Figure 1), for example, the data base BD can be continuously supplemented by new information, thus establishing the relation; treatment parameters = f(characteristics of the raw water//objectives) - This base BD,, which is initialised by the data derived from the knowledge possessed by the manufacturers of the performance of their equipment, thus becomes more reliable as operations progress and hence adapts to the specific qualities of the raw water of a given siter to the performance of the lamellar settler 27 (Figure 1) and to the equipment 24 for injecting reagents.

As stated above, an important aspect of the process according to the invention relates to the functioning 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 well the complexity of such a problem.

Presupposing a period which permits on average washing operations in 36hour phases, a first simple rule leads to a proposal for a period of 24 hours in the case of four filters and 48 hours in the case of the other two.

At time t=10 hoursi the history of the sludge loads and hydraulic loads admittedr and the forecasts of future loadsi make it possible to establish the probable durations before washing as set out in the table below:

Filter No. 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 peakpredicted for midday can be absorbed on five filters, so that it is possible to cut off the supply to one filter, and that filter No. 2 must be shut down up to 16 1 1 17 hours to enable it to be washed during a slack period. On the hypothesis of a shutdown of filter No. lp the following new prediction is established:

Filter No. 1 2 3 4 5 6 Duration(h) X 5 35 17 11 32 It is deduced from these values that filter No. 2 will have to be washed after 15 hours or shut down; but by applying the rules of the filtration unit it cannot be shut down without risking exceeding the maximum load during the peak. A new overall combination is therefore sought by setting an increase in yield for the physicochemical treatment installation and evaluating the consequences for the quality of the treatment and the costs. A new reasoning cycle is then begun and it can be seen that the complexity of this problem - which is not however the most complicated which has to be resolved justifies having recourse, in accordance with the invention, to an expert system.

Indeed, this system permits the purification potential of the biological purification installation to be determined on the basis of the degree of soiling of the filters employed and to incorporate the evaluation of the operating time of the latter as a function of the sludge loads and hydraulic loads which are applied.

More precisely. the biological filtration stage receives the residual flow of pollution after treatment by the physicochemical treatment installation. This flow is likewise characterised by a flow rate. an MIS concentration and a BOD concentrationj which are measured by sensors P2 and T2 (Figure 1) situated at the outlet of the preceding stage.

The operating parameters of each of the filters to achieve a given objective in terms of MIS and BOD at the outlet are thus the instantaneous velocity of 1 90/3181/01 18 the water through the filter, the latter being the result inter alia of the number of filters in use, and the flow of air to be injected, which is the result of a precise calculation on the basis of the pollution load.

In the system of the invention, each filter retains the evidence of its activity since the previous washing in the form of a degree of blocking and of a potential cycle duration before the next washing. It is obvious that these factors depend simultaneously on the MIS retained on the filter and on the BOD already eliminated by the biological activity. Experience likewise allows the incorporation, given the knowledge of the cycle durationt of the concept of profile of pollution received or to be received, which was the matter in question above.

This corresponds to the intuitive idea that the impact of a particular pollution load, corresponding to a given volume of water to be treated, on the filtration cycle duration or on the properties of the water treated is not the same, according to whether it is distributed equally over a constant flow or treated with peaks,. both hydraulic and of concentration.

In the same way as for the physicochemical stage, it is possible on the basis of the sensors situated on the water treated to continuously compare the results obtained and the results anticipated and to supplement the data base BD following certain, albeit relatively complex, rules so as to obtain the relations:

Treatment parameters = f(water treated.

objectives with respect to water filtered) At each instant#, each filter is thus characterised by its operating parameters and by a 90/3181/01 1 19 potential cycle duration in the light of the profiles of pollution to be received.

It is once more appropriate to note that the purification potential of the installation can also be determined on the basis of the quantity of biomass present in the activated sludge stage of the biological purification installation and that the expert system used within the framework of the present invention can incorporate the evaluation of the sludge 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 motor of the expert system can employ two types of reasoning, commonly called "forward chaining" and "backward chaining". With the first, reasoning is carried out using the rules on the basis of the known facts until the aim has been deduced. In backward chaining, in contrast, reasoning takes place starting from the aim, attempting to reduce it to equivalent secondary aims which are then the subject of an attempt at verification.

The expert system employed within the framework of the present invention can use these two types of reasoning: forward chaining when, on the basis of the quality of a batch of raw water, it defines the operating parameters of the physicochemical treatment installation; and backward chaining when, on the basis of the treatment objective, the state of the filters and the profile of pollution to be received, it defines the characteristics of the water subjected to settling which it ought to receive.

In conclusion, it can be seen that the regulation method according to the invention permits a continuous reaction to the actual effluents which enter a purification station, this being achieved as a function of the state of the latter and while 90/3181/01 4 1 k incornoratina time. This continuous adaptation opens the way in particular to various types of improvement in addition to those already desc ribed. Thus the quality of treatment is on the one hand improved by using each piece of equipment ppovided at all times to the best of its capabilities and, on the other hand, not all of the pieces of equipment provided are used at all times.

For a given treatment capacity, a superior overall reliability is obtained and the installation is given the capacity to mobilise either'additional physicochemical treatment or a certain number of filters which have until then been out of operationr to deal with an instantaneous pollution load which was unforeseen as compared with the profile types of pollution.

its functionina in c 90/3181/01 21

Claims (16)

1. A process for regulating installations for purifying waste waters these installations being of the type which employs a physicochemical treatment -and a biological treatment which can operate in parallel or in series, and of the type in which the residual flow of pollution is continuously influenced by varying at least one of the parameters comprising the quantity of flocculant added to the water to be treated, the distribution of the flow of water between the biological treatment and the physicochemical treatment. and the flow rate of air injected during the biological treatment, and, where applicable, the number of stages of treatment. characterised in that the variation of the parameters simultaneously takes account of the data comprising the assessments of the instantaneous pollution at the various levels of the purification treatmentr the forecasts of the pollution load over a period of at least 24 hours and the purification potential of the installation.

2. A process according to Claim 1, wherein the purification potential is determined on the basis of the soiling of at least one of the filters employed during the biological treatment and/or on the basis of the quantity of biomass present in the activated sludge stage.

3. A process according to Claim 1 or claim 2. wherein the forecasts of the value of the pollution load extend over a period which corresponds to a characteristic cycle of the pollution profile in terms of flow rate and content.

90/3181/01 1 22

4. A process according to Claim 3r wherein the forecasts of the value of the pollution load extend over a period of two times 24 hoursr and that they are examined after a time lapse of 48 hours.

-

5. A process according to any one of Claims 1 to 4, wherein the techniques of artificial intelligence are employed to influence the parameters on the basis of the data.

6. A process according to any one of Claims 1 to 5r further incorporating the evaluation of the operating time of the filters employed during the biological treatment as a function of the sludge loads and 15 hydraulic loads applied.

7. A process according to any one of Claims 1 to 67 further incorporating economic imperatives and, in particular, the cost of electrical energy as a 20 function of predetermined time segments.

8. A process according to any one of Claims 1 to 71 further incorporating the evaluation of the sludge load of an activated sludge reactor as a function of 25 the pollution already received or to be received.

9. A process according to any one of Claims 1 to 8p further incorporating the results previously obtained in order to improve its performance.

10. A process according to any one of Claims 1 to 9r wherein the parameters further comprise the stirring energy employed during the physicochemical treatment.

11. A process according to any one of Claims 1 to 10p wherein the pollution of the water to be treated 1 t R -90/3181/01 0 23 or undergoing treatment is assessed on the basis of at least one of the values of the matter in suspension (MIS)r the biological oxygen demand (BOD). the chemical oxygen demand (COD)r and the nitrogen and phosphorus contents.

12. A process according to Claim 11, wherein the assessment of the values MIS, BOD and COD, and of the nitrogen and phosphorus contents is effected on the basis of optical density measurements.

13. A process according to any one of Claims 1 to 12, wherein a unit for assessing the pollution comprising a lamellar settler, a UV/visible absorption spectrophotometer operating continuously on waste water and a microprocessor which processes the information received in accordance with a pre-established model is employed upstream of the physicochemical treatment.

14. A process according to any one of Claims 1 to 13, wherein a toxic substance detector, the data from which are incorporated to allow a correction by chemical means, and/or short circuiting of the biological treatment, and/or diversion of the toxic flow towards an intermediate storage reservoir, is employed upstream of the physicochemical treatment.

15. A process according to any one of Claims 1 to 14# wherein a device for assessing the pollution comprising a continuously operating UV/visible absorption spectrophotometer is employed upstream and/or downstream of the biological treatment.

16. A process according to any one of Claims I to 15y wherein a turbidimeter and/or a laser instrument for determining particle size is employed upstream 4, 24 and/or downstream of the physicochemical treatment and/or the biological treatment, for assessing the pollution of the water to be treated.

e co C. Ina tam Published 1988 at The Patent Office. State House. 6571 Hig'. Ho2borr.. London WC1R 4TP- Furth r pi s. y be ob ed from The Patent OfficeSales Branch, St Mary Cray. Orpington. Kent BR5 3RD- Printed by Multiplex techMqUes ltd. St Mary Cray. Kent. Con. 1/87.