FR3135548A1 - Method for optimizing the cleaning of a fluid circuit within an industrial installation - Google Patents

Abstract

Process for optimizing the cleaning of a fluid circuit within an industrial installation. The invention relates to a method for optimizing the cleaning of a fluid circuit within an industrial installation, said fluid circuit ensuring the implementation of a predetermined treatment process of a predetermined fluid in said industrial installation . Such a method comprises at least one iteration of the following steps: - adjustment (21), by a calculation machine (MC) of a configurable cleaning unit (UNP), of at least one parameter setting data of said cleaning unit configurable, depending at least on measurements delivered by a plurality of sensors (CPT) positioned on said fluidic circuit and characteristics intrinsic to said fluidic circuit, said predetermined fluid, and said predetermined treatment; - implementation (22) of optimized cleaning of said fluidic circuit, by said configurable cleaning unit, as a function of said at least one adjusted parameter setting data. Abstract Figure: Figure 2

Description

Method for optimizing the cleaning of a fluid circuit within an industrial installation

The invention lies in the field of cleaning fluid circuits within an industrial installation. The invention can in particular be implemented in the field of cleaning production lines for liquid or semi-liquid food products in the food industry.

Prior art

Many industrial installations include fluid circuits used for the implementation of treatment processes or production of a predetermined fluid. This is particularly the case for many companies in the agri-food sector, where fluid circuits are for example used for the conveying, processing and/or production of liquid or semi-liquid food products (milk, beer, ketchup , yogurt, etc.). Such fluid circuits are generally composed of a wide variety of structural elements, including for example storage tanks, equipment specific to the treatment to be carried out and/or to the fluid to be treated (coagulator, sterilizer, churn, pasteurizer, kettle, etc.), as well as piping systems (piping, valves, manifold, etc.) to convey the fluid between these different pieces of equipment.

In certain fields, and particularly in the agri-food sector which targets the production of foodstuffs, the cleaning of these fluid circuits is of capital importance, in particular to avoid any contamination and prevent any health risk. Also, solutions have been developed to ensure the implementation of regular cleaning (for example between each treatment cycle and/or production of the predetermined fluid) of a fluidic circuit, in the form of so-called “cleaning in place” (also called CIP systems, or CIP for “ Clean-in-Place ” in English), an example of which is presented schematically in relation to the . Such a cleaning system 12 comprises a set of tanks (121, 122, 123, 124) filled with different cleaning products (for example an alkaline detergent solution in the tank 121, an acid solution in the tank 122, water rinsing solution in the tank 123, a disinfectant solution in the tank 124, etc.) for the implementation of an automated sequence for cleaning a fluidic circuit 11. The automated cleaning sequence is generally predefined, and executed via a automaton which controls the cleaning system in place. As one or more cleaning cycles progress or at the end of one or more cleaning cycles, at least part of the cleaning products used is discharged to a treatment plant 13.

If it is certain that the automation thus implemented for cleaning the fluid circuits of an industrial installation presents undeniable advantages for obtaining certain guarantees both in terms of frequency and quality of cleaning, and that it thus contributes in particular to ensure the compliance of these installations on a regulatory level, the existing cleaning solutions in place nevertheless remain largely perfectible in several aspects.

In particular, it has been observed in numerous cases, notably through the study of the evolution of residual fouling within fluid circuits, that the results of cleaning carried out using existing cleaning-in-place solutions are often heterogeneous. It is therefore common for certain parts of the fluid circuit to suffer from a lack of cleaning, or at the very least seem to be less well cleaned than other parts of this circuit. One explanation is that current cleaning-in-place systems are generally configured manually by manufacturers, following very general configuration instructions provided by other manufacturers (i.e. equipment manufacturers), without real consideration of many specificities of the fluid circuit to be cleaned. In particular, these general instructions most often define configuration recommendations specific to each piece of equipment considered in isolation from other equipment in the fluidic circuit, and mainly boil down to proposals for adapting the configuration according to the nature of the fluid treated by the circuit. fluidics. In the agri-food sector in particular, the recommended configuration is thus essentially differentiated according to the “food recipe” implemented, for example depending on whether the fluid circuit is used to produce ketchup or a dairy product. These general instructions therefore ignore the specificity of the structure of the fluidic circuit considered, that is to say the way in which the different equipment which composes it is arranged (diameters of the pipes, distances between the equipment, volume of the tanks, impacts on the flow rate of fluid circulation, etc.). This results in the cleaning deficits mentioned above.

This lack of adaptation of the configuration of the cleaning system to the particularities of the fluid circuit to be cleaned as well as to other factors also has other undesirable consequences. Thus, in addition to the lack of cleaning observed on certain parts of a fluidic circuit, it has also been shown that other parts of the fluidic circuit are likely to be subject to excess cleaning with existing systems. cleaning in place. Such “overcleaning” is not desirable, in that it also raises significant problems, not only on an economic level (unnecessary quantities of cleaning products are used), but also on an ecological level (the products cleaning products once used are part of the polluting liquid discharges from the industrial installation). Studies have, for example, shown that 75% to 90% of waste from the dairy industries comes from cleaning operations on production lines. Beyond the health issue, there is therefore a strong ecological issue associated with the management of cleaning fluid circuits in industry.

It appears from the above that the cleaning of fluid circuits as it is currently implemented in many industrial installations is not always optimal. When they become aware of this, the manufacturers concerned can possibly call on specialist consultants to try to identify areas for modification of the configuration of the cleaning system, likely to improve the results. However, such a process can be long and costly, and is most often based on an empirical approach. In addition, the solutions identified in this framework generally lack adaptability, in the sense that any modification of the initial conditions is likely to call them into question, and therefore require a new iteration of the entire consultation process.

There is therefore a need for a solution to optimize the cleaning of fluid circuits within an industrial installation, which does not present the constraints of existing solutions.

The present technique makes it possible to propose a solution aimed at remedying certain disadvantages of the prior art. The present technique relates in fact to a method of optimizing the cleaning of a fluid circuit within an industrial installation, said fluid circuit ensuring the implementation of a predetermined treatment process of a predetermined fluid in said installation industrial. According to the general principle of the present technique, said method comprises at least one iteration of the following steps:

- adjustment, by a calculation machine of a configurable cleaning unit, of at least one parameter setting data of said configurable cleaning unit, based on at least measurements delivered by a plurality of sensors positioned on said fluidic circuit and intrinsic characteristics of said fluidic circuit, said predetermined fluid, and said predetermined treatment;

- implementation of optimized cleaning of said fluidic circuit, by said configurable cleaning unit, as a function of said at least one adjusted parameter setting data.

In a particular embodiment, said method comprises a phase of modeling clogging resulting from the implementation of the predetermined treatment process of the predetermined fluid within the fluidic circuit, and said step of adjusting at least one datum of parameterization takes into account said modeling of clogging.

According to a particular characteristic of this embodiment, said fouling modeling phase comprises the following steps:

- modeling of said fluidic circuit;

- modeling of said predetermined fluid;

- simulation of an implementation of said predetermined process, for said modeled fluid and by means of said modeled fluidic circuit, said simulation delivering said modeling of clogging.

According to a particular characteristic , the modeling of said predetermined fluid comprises the following steps:

- characterization of said predetermined fluid, comprising the association of a theoretical reference fluid to said predetermined fluid;

- obtaining a standard modeling of said theoretical reference fluid;

- adaptation of said modeling of said theoretical reference fluid, comprising a simplification of said theoretical reference fluid, delivering an adapted modeling corresponding to said modeling of the predetermined fluid.

According to a particular characteristic , the simplification of the modeling of the theoretical reference fluid includes the elimination, within the modeling of the theoretical reference fluid previously obtained, of molecules and/or compounds with low fouling power.

According to a particular characteristic , said adaptation further comprises the adjustment of at least one parameter of said standard modeling as a function of at least one characteristic associated with a particular class to which said predetermined fluid is attached.

According to another particular characteristic of this embodiment, said method comprises a phase of simulation of cleaning of said fluidic circuit implemented on the basis of said modeling of clogging, and said step of adjusting at least one parameter setting data holds takes into account the results of said cleaning simulation.

According to a particular characteristic, said cleaning simulation phase comprises the following steps:

- analysis of the results of said clogging modeling;

- determination of a plurality of cleaning strategies, based on the analysis of the results of said clogging modeling;

- for at least one cleaning strategy among said plurality of cleaning strategies, simulation of an implementation of cleaning of said clogging modeled by said cleaning unit parameterized as a function of the cleaning strategy considered;

- comparison of the results of said cleaning simulations, and selection of the cleaning strategy associated with the simulation leading to the greatest reduction in clogging for said clogging modeling.

In a particular embodiment, said method comprises at least one iteration of the following steps, subsequent to said step of adjusting at least one parameter setting item:

- implementation of a test cleaning by said configurable cleaning unit;

- analysis of residual clogging within the fluidic circuit, following said test cleaning;

- readjustment of the configuration of said configurable cleaning unit, depending on the results of the analysis of residual fouling.

According to another aspect, the proposed technique also relates to a device for optimizing the cleaning of a fluidic circuit within an industrial installation, said fluidic circuit ensuring the implementation of a predetermined treatment process of a predetermined fluid in said industrial installation. Such a device includes:

- means for adjusting at least one parameter setting data of a configurable cleaning unit, as a function of at least measurements delivered by a plurality of sensors positioned on said fluidic circuit and characteristics intrinsic to said fluidic circuit, said predetermined fluid , and said predetermined treatment;

- means for controlling said configurable cleaning unit, for implementing optimized cleaning of said fluidic circuit, as a function of said at least one adjusted parameter setting data.

According to another aspect, the proposed technique also relates to a computer program product downloadable from a communications network and/or stored on a computer readable medium and/or executable by a microprocessor, comprising program code instructions for the execution of a method of adding a communicating object to a wireless communication network as described above, when executed on a computer.

The proposed technique also relates to a computer-readable recording medium on which is recorded a computer program comprising program code instructions for the execution of the steps of the method as described above, in any of its embodiments.

Such a recording medium can be any entity or device capable of storing the program. For example, the medium may comprise a storage means, such as a ROM, for example a CD ROM or a microelectronic circuit ROM, or even a magnetic recording means, for example a USB key or a hard disk.

On the other hand, such a recording medium may be a transmissible medium such as an electrical or optical signal, which may be conveyed via an electrical or optical cable, by radio or by other means, so that the program computer it contains can be executed remotely. The program according to the invention can in particular be downloaded onto a network, for example the Internet network.

The different embodiments mentioned above can be combined with each other for the implementation of the invention.

Figures

Other characteristics and advantages of the invention will appear more clearly on reading the following description of a preferred embodiment, given by way of a simple illustrative and non-limiting example, and the appended drawings, among which:

presents a simplified diagram of a cleaning-in-place system according to the prior art;

presents a simplified diagram of a configurable cleaning unit for the implementation of a process for optimizing the cleaning of a fluidic circuit, in a particular embodiment of the proposed technique;

presents the different stages of a phase of modeling the clogging of a fluidic circuit, in a particular embodiment of the proposed technique;

shows an example of modeling a part of the structure of a fluidic circuit, in a particular embodiment of the proposed technique;

details the steps of modeling a fluid, in a particular embodiment of the proposed technique;

shows an example of modeling of clogging obtained for part of the structure of a fluidic circuit, in a particular embodiment of the proposed technique;

presents the different stages of a simulation phase of cleaning a fluidic circuit, in a particular embodiment of the proposed technique;

presents a simplified architectural example of a device for optimizing the cleaning of a fluidic circuit, in a particular embodiment of the proposed technique.

Detailed description of the invention

The proposed technique makes it possible to remedy some of the aforementioned drawbacks.

The proposed technique relates more particularly to a process making it possible to optimize the cleaning of a fluid circuit within an industrial installation. The proposed technique is for example particularly suitable for use in the food industry, to optimize the cleaning of fluid circuits used for the production of food liquids or semi-liquids such as the production of milk, beer, tomato sauce, etc. The preceding example is however given for purely illustrative and non-limiting purposes, the proposed technique also being suitable for use in other types of industries, for example the cosmetics and pharmaceutical industries.

As described below, the present technique offers multiple advantages over existing cleaning-in-place systems, in that it allows:

- wash better, which is of hygienic interest, with a reduction in the risk of contamination;

- reduce the quantities of cleaning products used and reduce discharges, which is of ecological interest and makes it possible to reduce the carbon footprint of the industries concerned.

In all the figures in this document, elements of a similar nature are designated by the same numerical reference.

The general principle of a process for optimizing the cleaning of a fluid circuit according to the present technique is illustrated in relation to the , in a particular embodiment. Conventionally, an automated cleaning sequence SEQ of the fluidic circuit CF is executed using different cleaning products available in the tanks of a configurable cleaning unit UNP, during a cleaning phase generally implemented after a phase production, that is to say after use of the fluidic circuit CF for the implementation of a predetermined treatment process TT of a predetermined fluid FL. Such a predetermined treatment TT can for example take the form of a transformation of a fluid into another fluid, of a production of a fluid, and/or of the conveyance of a fluid from a point to a other of an industrial installation. The automated cleaning sequence SEQ is controlled by a calculation machine MC of the configurable cleaning unit UNP, on the basis of a set of parameterization data PRM. The automated cleaning sequence SEQ more particularly comprises a succession of cleaning steps implemented in a given order, the nature and order of these steps being likely to vary depending on the characteristics of the FL fluid treated in the production phase by the fluidic circuit CF, and the predetermined treatment TT implemented by this circuit during this production phase. In the non-limiting example illustrated in relation to the , the automated cleaning sequence SEQ includes the following main steps: pre-wash PL, soda wash LS, first rinse R1, acid wash LA, second rinse R2, disinfection DF, third rinse R3.

According to the general principle of the present technique, a process of continuous optimization of the cleaning of the fluidic circuit is implemented.

To this end, the present technique relates, in one of its aspects, to a process for optimizing the cleaning of the fluidic circuit comprising at least one iteration of the following steps:

- a step of adjustment 21, by the calculation machine MC of the configurable cleaning unit UNP, of at least one parameter setting data of said configurable cleaning unit, as a function not only of intrinsic characteristics of the fluidic circuit to be cleaned , characteristics intrinsic to the predetermined fluid treated, and characteristics intrinsic to the predetermined treatment carried out, but also as a function of measurements delivered by a plurality of CPT sensors positioned on said fluid circuit;

- a step 22 of implementing optimized cleaning of the fluidic circuit, by the configurable cleaning unit UNP, as a function of said at least one parameter setting data adjusted in step 21.

In this way, the proposed technique makes it possible in particular to guarantee automated, intelligent, and real-time balancing of the four main interdependent factors known to contribute together to the effectiveness of cleaning, namely: Action Time, Mechanical action, Concentration (or Chemistry, this parameter having as its object the choice and dosage of the products used for cleaning), and the Cleaning Temperature.

Indeed, according to the proposed technique, these four factors – often referred to in the literature under the acronym “TACT” – are automatically analyzed by the MC calculation machine, in particular via measurements reported by a set of CPT sensors positioned on the entire CF fluidic circuit to be cleaned (i.e. on at least the majority of specific equipment present in the fluidic circuit). Such sensors include, for example, temperature, pressure, flow sensors, etc. More particularly, the data reported from the CPT sensors are processed by the MC calculation machine, in addition to other data, examples of which are detailed below, so that a failure observed on one of the TACT factors ( e.g. a cleaning temperature that is too low, an action time that is too short, insufficient mechanical action or an unsuitable dosage) either automatically corrected (for example by increasing the temperature if it is too low) or compensated automatically via a compensatory adjustment of a TACT factor other than the faulty factor (for example by compensating a mechanical action identified as too limited due to insufficient flow rate with a longer action time). The calculation machine MC thus operates a continuous adjustment of the PRM parameterization data of the configurable cleaning unit UNP (for example by adapting the temperature, the volume of cleaning product sent into the fluidic circuit, the number of times or the product of cleaning runs through the entire fluidic circuit i.e. the number of loops, etc.), so as to optimize the cleaning of the fluidic circuit, i.e. to ensure that all parts of the fluidic circuit are sufficiently cleaned , but none of these parts are subject to excessive cleaning.

According to the proposed technique, other data than just the data from the CPT sensors are also taken into account for the implementation of step 21 of adjusting at least one parameter setting data of the configurable cleaning unit, in order to allow the realization of the TACT compensation mechanism previously presented. Such data are for example available within data structures accessible from the calculation machine MC, and include at least:

- data intrinsic to the fluidic circuit CF (such data typically includes information relating to the structure and arrangement of the fluidic circuit, for example in the form of a virtual division into sections of this circuit and information associated with each of said sections);

- data intrinsic to the predetermined TT treatment implemented (for example data relating to the treatment temperature(s), the duration of a production cycle, etc.);

- data intrinsic to the predetermined fluid FL treated (e.g. composition, rheology, viscosity, dry matter, fat, etc.).

In a particular implementation presented in detail later in this document, the step of adjusting at least one parameter setting data of the configurable cleaning unit also takes into account data from the modeling phase and/or simulations.

According to various particular embodiments, other complementary data (coming for example from other data sources AS) can optionally also be provided as input to the calculation machine MC, for the implementation of step 21 d 'adjustment of at least one parameter setting data of the configurable cleaning unit, such as for example:

- data relating to the cleaning programs configured by default (e.g. temperature, time, conductivity, flow rate, chemical product consumption, energy consumption, etc.) and/or the cleaning step (PL, LS, R1, LA, etc.) in progress;

- data relating to histories of microbiological results and/or non-compliance results associated with the predefined treatment implemented by means of the fluidic circuit;

- data from wastewater treatment plants (e.g. biochemical oxygen demand BOD, chemical oxygen demand COD, hydrogen potential pH, volumes, etc.);

- regulatory data and data relating to the microbiological limits authorized for each treatment implemented and/or product used.

In a particular embodiment, the steps 21 of adjusting at least one parameter setting data of the configurable cleaning unit and 22 of adapting the cleaning sequence according to said at least one adjusted parameter setting data are implemented multiple times during the automated SEQ cleaning sequence. They are for example implemented at least once before or during each of the main stages (PL, LS, R1, LA, etc.) forming the automated cleaning sequence 21, thus making it possible to carry out continuous optimization or at any time. the least quasi-continuous of the cleaning process.

According to a particular characteristic, the implementation of step 21 of adjusting at least one parameter setting data of the configurable cleaning unit is based on an artificial intelligence brick. To this end, solutions based on machine learning techniques and on the use of one or more artificial neural networks within the calculation machine MC are for example implemented, these neural networks taking as input all or part of the input data described above (at a minimum the measurements from the CPT sensors, and the intrinsic characteristics of the fluidic circuit, the predetermined fluid, and the predetermined treatment) to deliver at the output adjusted PRM parameterization data making it possible to configure the configurable cleaning unit UNP so as to obtain optimized cleaning at the instant considered (and for example particularly adapted to the current step of the cleaning sequence SEQ).

We now present, in the remainder of the document, a particular implementation of the technique previously described, in which data from preliminary modeling and simulation phases (represented in dotted lines on the ) are used to further improve the optimization of the cleaning of a fluid circuit within an industrial installation. More particularly, in this particular implementation, it is proposed to implement a phase of modeling the clogging of the fluidic circuit (on the basis of characteristics taking into account in particular the structure of the fluidic circuit, the predetermined treatment process implemented by this circuit, and the predetermined fluid object of this treatment), then to exploit the modeling of clogging obtained MDL-E by injecting it into the artificial intelligence brick implemented by the calculation machine MC, in order to either serve additional input parameter used directly and as such by the calculation machine MC to further improve the optimization of cleaning, or to serve as a basis for the calculation machine to implement a cleaning simulation phase allowing to test different cleaning strategies of the fluidic circuit, and to use the SIM-N results of such a simulation to further improve the cleaning optimization.

There illustrates a phase of modeling clogging as briefly presented previously, which can be integrated into a process for optimizing the cleaning of a fluidic circuit in a particular embodiment of the proposed technique. Such a clogging modeling phase, which makes it possible to estimate the clogging resulting from the implementation of the predetermined treatment process of the predetermined fluid within the fluidic circuit, can be implemented by the calculation machine MC itself, or by one or more third-party calculation machines.

This phase of modeling clogging initially comprises a step 31 of modeling the fluidic circuit CF and a step 32 of modeling the predetermined fluid FL. A simulation of the implementation of the predetermined process, for the fluid modeled in step 32 and by means of the fluidic circuit modeled in step 31, is then carried out in a step 33, said simulation delivering an MDL clogging modeling -E.

The modeling 31 of the fluidic circuit includes for example a two-dimensional, ideally three-dimensional, modeling of the entire fluidic circuit. An example of the result of a three-dimensional modeling of a particular structural element of a fluidic circuit is illustrated in relation to the .

In a particular embodiment, illustrated in relation to the , step 32 of modeling the predetermined fluid FL comprises:

- a step 321 of characterizing the predetermined fluid FL, comprising the association of a theoretical reference fluid to said predetermined fluid;

- a step 322 of obtaining a standard modeling of said theoretical reference fluid;

- a step 323 of adaptation of said modeling of said theoretical reference fluid, delivering an adapted modeling corresponding to said modeling of the predetermined fluid.

Step 321 of characterizing the predetermined fluid aims to associate it with a theoretical reference fluid supposed to have similar characteristics (in particular for example in terms of its chemical composition, its rheology, etc.), and for which a standard modeling has already been saved in the system (for example in a memory of the configurable cleaning unit). By way of illustration, if the predetermined fluid treated in the production phase by the fluidic circuit is goat's milk, the theoretical reference fluid is for example cow's milk, relatively close but more common, and which has already as such the subject of a standard modeling accessible by the configurable cleaning unit.

This standard modeling of the theoretical reference fluid is obtained in step 322.

We then proceed, in step 323, to an adaptation of this standard modeling of the theoretical reference fluid. Such an adaptation includes at least one step of simplification of this standard model, as well as possibly various adjustments of certain parameters of this model, as described below.

According to a particular characteristic, the simplification implemented comprises the elimination, within the modeling of the theoretical reference fluid previously obtained in step 322, of the elements (molecules and/or compounds) of which a prior analysis has shown that they contribute little or that they do not contribute to the clogging observed within a fluidic circuit. In other words, to carry out such a simplification, we mainly retain two types of elements in the adapted modeling:

- on the one hand the elements identified as contributing directly to fouling, due to their significant intrinsic fouling power (studies have for example shown that among all the compounds present in milk, numbering several hundred, two molecules only are the origin of the majority of direct fouling: beta-lactoglobulin and calcium);

- on the other hand, elements identified as contributing indirectly, that is to say elements with low fouling power or less fouling power, but which find themselves in practice trapped in the fouling generated by other elements , and which therefore increase this fouling by accumulation.

Elements that do not fit into one or the other of these two types are thus removed from the model, which greatly reduces its complexity.

In addition to this simplification, adjustments of various parameters of the standard modeling of the theoretical reference fluid are possibly also made, in order to obtain a modeling of the predetermined fluid which is the most coherent and the most representative possible from the perspective of a fouling simulation. Such adjustments may in particular take into account characteristics associated with a particular class to which the predetermined fluid belongs, in particular characteristics relating to the composition of the fluid (for example via an adjustment of parameters defining the concentration of the various constituent elements retained in the model simplified) or the rheological characteristics of the fluid (for example via an adjustment of parameters relating to the hydrogen or pH potential, thermal and/or electrical conductivity, viscosity, etc.). Such adjustments can for example be implemented using reference data – possibly accessible in the form of characteristic tables or specification charts – available for different classes of a particular type of fluid (for example, if we considers the "milk" type fluid, such reference data can list the main differences between the classes of milk "cow's milk", "goat's milk", "sheep's milk", or even more precise classes "cow's milk". mountain cow”, “milk from cows raised in the plains”, etc.).

Once the modeling of the fluidic circuit and the predetermined fluid has been carried out, step 33 of fouling simulation is implemented. Such a simulation is based for example on techniques called “computational fluid mechanics” (or MFN, more commonly referred to by the English term “ C omputational Fluid Dynamics ” or CFD) which consists of studying the movements of a fluid, or their effects, by the numerical resolution of the equations governing the fluid. The simulation is obviously carried out on the basis of the modeling of the fluidic circuit and the modeling of the predetermined fluid, but also takes into account parameters linked to the predetermined treatment process TT to be simulated (this treatment process being for the record the reason for being of the fluidic circuit, ie what it is used for), including for example treatment temperature parameters, duration of production cycles, etc. The number of parameters coming into play in such a simulation being potentially very important, we therefore understand the interest in carrying out this simulation on the basis of a modeling of the predetermined fluid which results from a simplification of a theoretical model. standard, as described previously in relation to step 32: such simplification makes it possible to drastically reduce the complexity of the calculations to be carried out within the framework of such a simulation, and therefore to reduce the simulation times and/or the power of calculation required, while ultimately retaining a correct approximation of the clogging. The MDL-E clogging modeling obtained at the end of this simulation includes an estimate of the mass and nature of the clogging, as well as an estimate of the locations of these clogging within the fluidic circuit, as illustrated for example in relation to the in a particular embodiment.

According to the present technique, the modeling of MDL-E fouling thus obtained can be exploited in different ways.

In a particular embodiment, it can be injected directly into the input of the calculation machine MC, as additional input data of the artificial intelligence brick (for example of one or more neural networks pre-trained to take taking into account, in addition to other input data, such modeling of clogging for the output delivery of an optimized configuration of the configurable cleaning unit).

In a particular alternative or complementary embodiment, modeling of clogging can also be used to test different cleaning strategies, by digital simulation. In this case, the method of optimizing the cleaning of a fluidic circuit then comprises a cleaning simulation phase comprising the following steps illustrated in relation to the :

- analysis 41 of the results of MDL-E fouling modeling;

- the determination 42 of a plurality of cleaning strategies, based on the analysis of the results of said clogging modeling;

- for each cleaning strategy among said plurality of cleaning strategies, the simulation 43 of an implementation of cleaning of said clogging modeled by said cleaning unit parameterized as a function of the cleaning strategy considered;

- comparison 44 of the results of said cleaning simulations, and selection of the cleaning strategy associated with the SIM_N simulation leading to the greatest reduction in clogging for said clogging modeling.

The MC calculation machine can then use the results of this cleaning simulation to adjust at least the settings of the cleaning unit, depending for example on the selected cleaning strategy.

Similar to the fouling modeling phase, the simulation phase can be implemented via computational fluid dynamics (CFD) techniques.

According to the present technique, in a particular embodiment, parameter adjustments are also implemented subsequent to the parameterization of the configurable cleaning unit. More particularly, after the implementation, following a production phase, of a test cleaning of the fluidic circuit carried out in situ and in real conditions via the configurable cleaning unit previously configured in accordance with any of the modes of embodiment described above, an analysis is carried out of the actual residual clogging observed within the fluidic circuit, then a readjustment of at least one parameter setting data of the cleaning unit according to the results of this analysis. In a continuous optimization approach, several iterations of this readjustment process via in-situ tests and in real conditions are carried out, until an optimal configuration of the cleaning unit is identified, i.e. the best compromise between quality cleaning and control of discharges.

According to another aspect, the proposed technique also relates to a device for optimizing the cleaning of a fluidic circuit within an industrial installation, said fluidic circuit ensuring the implementation of a predetermined treatment process of a predetermined fluid in said industrial installation. More particularly, such a device comprises:

- means for adjusting at least one parameter setting data of a configurable cleaning unit, as a function of at least measurements delivered by a plurality of sensors positioned on said fluidic circuit and characteristics intrinsic to said fluidic circuit, said predetermined fluid , and said predetermined treatment;

- means for controlling said configurable cleaning unit, for implementing optimized cleaning of said fluidic circuit, as a function of said at least one adjusted parameter setting data.

There represents, in a schematic and simplified manner, in a particular embodiment, the structure of such a device, which generally corresponds to a machine for calculating a configurable cleaning unit.

For example, the device according to the proposed technique comprises a memory 81 consisting of a buffer memory M, a processing unit 82, equipped for example with at least one microprocessor μP, and controlled by the computer program Pg 83, putting implement the method for optimizing the cleaning of a fluid circuit according to the invention.

Upon initialization, the code instructions of the computer program 83 are loaded into the buffer memory before being executed by the processor of the processing unit 82. The processing unit 82 receives as input E for example data specific to the fluidic circuit considered, to the predetermined treatment process implemented by this circuit, and to the predetermined fluid subject to this treatment, data reported by sensors placed on the fluidic circuit, and possibly data from modeling phases and /or simulation previously implemented. Other additional data available for example within third-party data structures accessible to the cleaning optimization device are possibly also provided as input E.

The microprocessor of the processing unit 82 then carries out the steps of the method of optimizing the cleaning of a fluidic circuit, according to the instructions of the computer program 83, to provide for example at output S at least one adjusted parameter setting data a configurable cleaning unit, for implementing optimized cleaning of the fluidic circuit. These operations can for example be implemented by means of an artificial intelligence brick of the processing unit 82, for example by means of one or more neural networks pre-trained for this objective.

According to another aspect, the proposed technique also relates to a computer program product comprising computer-executable program code recorded on a non-transitory computer-readable storage medium, the computer-executable program code, when executed, performing the method of optimizing the cleaning of a proposed fluidic circuit. The computer program product may be recorded on a CD, hard disk, flash memory or other suitable computer (or processor) readable media. It can also be downloaded from the Internet and installed in a device so as to allow the implementation of a process for optimizing the cleaning of a fluidic circuit, in any of the embodiments previously described.

Claims (11)

Method for optimizing the cleaning of a fluidic circuit (CF) within an industrial installation, said fluidic circuit ensuring the implementation of a predetermined treatment process (TT) of a predetermined fluid (FL) in said industrial installation, said method being characterized in that it comprises at least one iteration of the following steps:
- adjustment (21), by a calculation machine (MC) of a configurable cleaning unit (UNP), of at least one parameter setting data of said configurable cleaning unit, based on at least measurements delivered by a plurality sensors (CPT) positioned on said fluidic circuit and intrinsic characteristics of said fluidic circuit, said predetermined fluid, and said predetermined treatment;
- implementation (22) of optimized cleaning of said fluidic circuit, by said configurable cleaning unit, as a function of said at least one adjusted parameter setting data. Method according to claim 1, said method being characterized in that it comprises a phase of modeling clogging resulting from the implementation of the predetermined treatment process of the predetermined fluid within the fluidic circuit, and in that said step d The adjustment of at least one parameter setting data takes into account said clogging modeling (MDL-E). Method according to claim 2, characterized in that said fouling modeling phase comprises the following steps:
- modeling (31) of said fluidic circuit;
- modeling (32) of said predetermined fluid;
- simulation (33) of an implementation of said predetermined process, for said modeled fluid and by means of said modeled fluidic circuit, said simulation delivering said modeling of clogging (MDL-E). Method according to claim 3, characterized in that the modeling of said predetermined fluid comprises the following steps:
- characterization (321) of said predetermined fluid, comprising the association of a theoretical reference fluid to said predetermined fluid;
- obtaining (322) a standard modeling of said theoretical reference fluid;
- adaptation (323) of said modeling of said theoretical reference fluid, comprising a simplification of said theoretical reference fluid, delivering an adapted modeling corresponding to said modeling of the predetermined fluid. Method according to claim 4, characterized in that the simplification of the modeling of the theoretical reference fluid comprises the elimination, within the modeling of the theoretical reference fluid previously obtained, of molecules and/or compounds with low fouling power. Method according to claim 4, characterized in that said adaptation further comprises the adjustment of at least one parameter of said standard modeling as a function of at least one characteristic associated with a particular class to which said predetermined fluid is attached. Method according to claim 2, said method being characterized in that it comprises a phase of simulation of cleaning of said fluidic circuit implemented on the basis of said modeling of clogging (MDL-E), and in that said step of adjustment of at least one parameter setting data takes into account the results of said cleaning simulation (SIM-N). Method according to claim 7, characterized in that said cleaning simulation phase comprises the following steps:
- analysis (41) of results of said clogging modeling (MDL-E);
- determination (42) of a plurality of cleaning strategies, based on the analysis of the results of said clogging modeling;
- for each cleaning strategy among said plurality of cleaning strategies, simulation (43) of an implementation of cleaning of said clogging modeled by said cleaning unit parameterized as a function of the cleaning strategy considered;
- comparison of the results of said cleaning simulations, and selection (44) of the cleaning strategy associated with the simulation leading to the greatest reduction in clogging for said clogging modeling. Method according to claim 1, characterized in that it comprises at least one iteration of the following steps, subsequent to said step of adjusting at least one parameter setting item:
- implementation of a test cleaning by said configurable cleaning unit;
- analysis of residual clogging within the fluidic circuit, following said test cleaning;
- readjustment of the configuration of said configurable cleaning unit, depending on the results of the analysis of residual fouling. Device for optimizing the cleaning of a fluidic circuit within an industrial installation, said fluidic circuit ensuring the implementation of a predetermined treatment process of a predetermined fluid in said industrial installation, said device being characterized in that which it includes:
- means for adjusting at least one parameter setting data of a configurable cleaning unit, as a function of at least measurements delivered by a plurality of sensors positioned on said fluidic circuit and characteristics intrinsic to said fluidic circuit, said predetermined fluid , and said predetermined treatment;
- means for controlling said configurable cleaning unit, for implementing optimized cleaning of said fluidic circuit, as a function of said at least one adjusted parameter setting data. Computer program product downloadable from a communications network and/or stored on a computer-readable medium and/or executable by a microprocessor, characterized in that it comprises program code instructions for executing a method for optimizing the cleaning of a fluidic circuit within an industrial installation according to any one of claims 1 to 9, when executed on a computer.