WO2019081252A1

WO2019081252A1 - Acidogenic fermentation process for the production of organic acids having from 2 to at least 5 carbon atoms and corresponding facility

Info

Publication number: WO2019081252A1
Application number: PCT/EP2018/078066
Authority: WO
Inventors: Maria Das Dores TAVARES GURGO E CIRNE; Sofiane MAZEGHRANE; Sébastien LACROIX
Original assignee: Veolia Environnement
Priority date: 2017-10-26
Filing date: 2018-10-15
Publication date: 2019-05-02
Also published as: FR3072971B1; FR3072971A1

Abstract

A process, and corresponding facility, for acidogenic fermentation of a fermentable substrate by a microbial consortium in a single-stage reactor, which is intended to provide at least one fluid effluent comprising a mixture of organic acids having from two to at least five carbon atoms and a gaseous effluent. The process comprises a step of advanced control of the acidogenic fermentation carried out, by virtue of a control-order system which makes it possible to maintain throughout the advanced control step in the reactor: the volumetric loading of organic matter, the hydraulic residence time, the residence time of the particulate matter, the temperature and the pH, in predetermined value ranges.

Description

Acidogenic fermentation process for the production of organic acids of 2 to at least 5 carbon atoms and corresponding plant

1. Field of the invention

The field of the invention is that of acidogenic fermentation processes.

More specifically, the invention relates to an improved acidogenic fermentation process for the industrial production of chemical platform molecules, especially for the production of organic molecules having from two to at least five carbon atoms.

In particular, the invention relates to an improved acidogenic fermentation process for the industrial production of carboxylic acids having from two to at least five carbon atoms.

The invention also relates to corresponding installations. 2. Prior Art

The awareness of the limited availability of petroleum resources is reviving the interest of biomass as a resource for the production of molecules of industrial and energy interest, as well as the use of industrial biotechnology to ensure the conversion of this raw material into products. interest.

Among the industrial biotechnologies, the acidogenic fermentation allows fermentable organic carbon to convert fermentable organic carbon into different products, these various products notably comprising carboxylic acids, in particular volatile fatty acids (VFAs) and lactic acid, alcohols. , dihydrogen and carbon dioxide. The production of carboxylic acids, and by extension the other products of acidogenic fermentation, is also known as the "carboxylic platform" (see Adgler et al., 2011), "Waste to bioproduct conversion with undefined cultures: the carboxylate platform ").

Acidogenic / acetogenic fermentation can be followed by a methanogenic step in an anaerobic digestion process, thus producing carbon dioxide. carbon and methane. Compared to the gaseous products resulting from the acidogenic fermentation (CH ₄ , H ₂ ), the volatile fatty acids (VFAs) of the carboxylic platform have many advantages: they have a high market value, a high energy density and have a high capacity of storage and transport. AGVs are the organic acids most often sought during acid-proof fermentation because of their various applications. In particular, they can be used as intermediates for methanation or as substrates for the production of biopolymers (PHA). Compared to other molecules such as succinic acid, butanol and acetone, which are produced using pure bacterial cultures, their production conditions are generally less demanding and expensive.

It has been described in the prior art, in particular in the document US 2015/0191754, an acidogenic fermentation process using a pure strain of bacteria with a mixture of sludge and kitchen waste. This process requires two distinct steps: a first step of pre-fermentation in a reactor to produce propionic acid precursors, the alkalinity being controlled during this step via the addition of calcium hydroxide. A second step, in which the reactor discharge is centrifuged, sterilized, and then sent to an acidogenic fermentation tank in which a culture of Propionibacterium is inoculated. Although this method makes it possible to obtain a reasonable concentration of propionic acid, between 9.5 and 16.8 gDCO / L, in the final fermentation juice, it nevertheless has several disadvantages. Indeed, the method can be implemented with a low diversity of substrates and requires a step of significant pretreatment of these substrates. The use of a pure strain of bacteria also has a high cost and makes the process particularly tedious to implement, because the process comprises a costly sterilization step and the fermentation must be carried out under axenic conditions. Finally, a pure strain of bacteria ensures a very incomplete conversion of substrates to complex compositions. To summarize, the use of a pure culture of bacteria in an acidogenic process makes it possible to achieve a high selectivity of the products derived from fermentation but in return requires an expensive and tedious implementation of the process.

An alternative to using a pure strain of bacteria is the use of mixed microbial cultures (microbial consortium). However, the industrial exploitation of acidogenesis by mixed cultures of bacteria raises important challenges both in terms of fermentation, and in particular the control of the spectrum of products formed, than the separation and recovery of these products. It is for example known from: Bengtsson et al. (2008); "Acidogenic fermentation of industrial wastewater: effects of chemostat retention time and pH on volatile acids productions", an acidogenic fermentation process using a mixed culture of bacteria with paper mill effluents. Experiments have shown that at a temperature of 37 ° C, a maximum concentration of volatile fatty acids of 5 to 6 gDCO / L could be obtained. This maximum concentration value is relatively low and can partly be explained by the phenomenon of inhibition (product toxicity) induced by certain acids which, at high concentration, block the reaction process of acidogenic fermentation. In addition, Bengtsson et al. does not describe means for controlled production and targeting of the volatile fatty acid spectrum from the process. A key point in the development of the carboxylic platform from fermentation is to develop techniques to reach high concentrations of carboxylic acids, and more particularly volatile fatty acids, during the acidogenic fermentation because their concentration strongly impacts the cost of product recovery step. The limitation of production by inhibition phenomenon of carboxylic acids (AGVs, lactate, etc.) can be removed by using specific in-situ extraction systems. The main research has been carried out to date on pure cultures of lactic acid and succinic acid production (see Arslan et al (2016): Selective short chain carboxylates production: a review on control mechanisms to direct mixed fermentations culture "). For example, it is possible to consider controlled extraction of volatile fatty acids to non-inhibitory levels by applying the principles described in the patent application WO 2011/055092, filed by the Applicant. However, the effectiveness of this method has not been demonstrated on fermentation processes producing high concentrations of organic acids.

The document WO 2016/012701 describes a process for producing organic acids having 1 to 8 carbon atoms from a fermentable biomass. Extraction of AGVs is implemented to minimize the presence of inhibitory concentration of carboxylic acids. The process generally makes it possible to obtain AGVs at a concentration of between 10 g / l and 20 g / l. This method, however, has several disadvantages. In addition to the need to extract the organic acids to avoid the effects of inhibition by accumulation, the main fermentation used in this process is carried out by microorganisms from a well-defined ecosystem, eukaryotic or prokaryotic, such as bacteria, fungi, algae or yeasts. The organic acids produced then comprise mainly between 2 to 4 carbon atoms. To obtain organic acids of longer chain, including organic acids comprising 6 carbon atoms, a secondary fermentation is necessary (chain elongation). This process may require the establishment of several reactors arranged in series. The continuous measurement of the concentration of organic acids is a key point to guarantee the performance of the carboxylic platform. However, this continuous measurement remains a challenge to date because of the sensors, which are the weak link in the real-time control chain of the fermentation and production process of organic acids. This observation has been known for three decades but has never been the subject of developments on an industrial scale. The concentration and distribution of organic acids, key variables to drive system performance can only be measured using equipment, whose industrial use is at the research stage: automated titration, online chromatography, spectrometry, etc. Another notable weakness of previous work is the lack of coordination between different simple indicators (pH sensors, potential oxidation-reduction, biogas composition, etc.) to translate the performance of the organic acid production process.

Moreover, there is little or no work focusing on real-time control and optimization of organic acid production. Indeed, most controllers have been developed to minimize the production of organic acids and to avoid acidification of an anaerobic digestion reactor. By regulating (limiting) the production of AGVs, it is then possible to maintain environmental conditions conducive to the growth of methanogenic archaea. These objectives have been described in US Pat. No. 6,551,510 B1. However, they are in contradiction with the objectives of the current patent application.

Finally, processes such as that described in WO 2007/009085 make it possible to obtain high concentrations of AGVs, but the corresponding processes and installations are expensive and complex to implement. Indeed, WO 2007/009085 describes a four-stage reactor.

3. Objectives of the invention

The object of the invention is to propose an improved acidogenic fermentation process which makes it possible to overcome at least certain disadvantages and limitations of the prior art.

The object of the invention is in particular to propose an improved process making it possible to achieve both high productivity, high yield and high selectivity of organic acids having from two to at least five carbon atoms, while keeping energy costs and low production.

The invention also aims to provide a stable and robust process despite disturbances (pH, inhibition, toxicity, etc.) and / or variations in the quality (composition of organic materials) and / or the amount of the feedstock. fermentable organic carbon. The invention also aims, at least according to some embodiments of the invention, to provide a method comprising an in-situ estimate of the concentration and nature of the organic acids.

4. Presentation of the invention

The present invention relates to a process for the acidogenic fermentation of a fermentable substrate by a microbial consortium in a single-stage reactor, which is intended to provide at least one fluid effluent comprising a mixture of organic acids having from two to at least five atoms of carbon and a gaseous effluent. The method comprises an advanced stage of acidogen fermentation conducted using a control-command system that allows maintenance during the advanced driving step in the single-stage reactor: the volume load material organic, hydraulic residence time (HRT), particulate matter residence time (SRT), temperature and pH, within predetermined ranges of values.

According to wide ranges of predetermined values, the volume load of organic material is maintained at a value between 10 and 80 gDCO / L / day, the hydraulic residence time at a value between 0.5 days and 4 days, the time of residence of the particulate matter at a value between 0.5 days and 7 days, the temperature at a value between 20 ° C and 40 ° C and the pH at a value between 5 and 7.

According to smaller ranges of values, the volume load of organic matter is maintained at a value between 30 and 60 gDCO / L / day, the hydraulic residence time at a value between 1 day and 3 days, the residence time particulate matter between 1 day and 4 days, the temperature between 30 ° C and 37 ° C and the pH between 5.5 and 6.5.

The organic acids of the organic acid mixture of the fluid effluent resulting from the process of the present invention preferably comprise from two to five carbon atoms and very preferably from two to six carbon atoms. carbon.

This improved acidogenic fermentation process provides controlled production in terms of concentration, composition and productivity of organic acids having from two to at least five carbon atoms using mixed microbial cultures. Controlled production is guaranteed by real-time optimization of the production of organic molecules using the control-command system to influence the operating conditions and the specialization of microbial biomass.

"Organic acids" are organic compounds having acidic properties in aqueous media. Carboxylic acids are a particular subclass of the organic acids preferentially produced by the process according to the invention. Among this subclass of carboxylic acids, the process according to the invention makes it possible in particular to produce volatile fatty acids with the presence or absence of lactic acid. The term "volatile fatty acid", also known by the French acronym AGV or the acronym VFA, means an aliphatic chain carboxylic acid having from 2 to 6 carbon atoms.

The term "volume load of organic matter" or "organic load" or the acronym "OLR" means the mass of substrate characterized by the Chemical Oxygen Demand (COD), introduced per day in a certain volume of reactor. It can be expressed in grams of COD (gDCO) per liter of reactor per day. The determination of the COD can be made according to the NF T90-101 standard.

The term "hydraulic residence time" or the acronym "HRT" means the time during which the liquid fraction of the substrate (liquid fraction of the feed) remains in the reactor (in contact with the microbial consortium). It represents the ratio between the reactor volume and the substrate feed rate. It is expressed in days.

"Particulate matter residence time" or the acronym "SRT" refers to the length of time that particulate matter, including the microbial consortium, is in the reactor. It represents the ratio between the amount of particulate matter in the reactor and the amount of particulate matter drawn off per day. Depending on the configuration of the facility and the nature of the microbial consortium and fermentable substrate, the TRS and HRT values may be equal or decoupled. TRS is also expressed in days.

By "productivity" is meant the mass flow rate of organic acids having from two to at least five carbon atoms produced by the process per unit reactor volume. It is usually expressed in g ( _AC ides ₎ / l- / day. In the case where the production of AGVs is targeted in particular, it is expressed in g ( _A Gv ₎ / L / day. According to the embodiment of the invention, the productivity can in particular be expressed for organic acids / AGVs having two to five carbon atoms or organic acids / AGVs having two to six carbon atoms.

By conversion "yield" is meant the ratio of the mass concentration of organic acids having from two to at least five carbon atoms produced converted to COD on the mass concentration (COD) of introduced fermentable substrate (gAcid-Dco / gDco ) - In the case where the production of AGVs is targeted in particular, the conversion efficiency corresponds to (gAGv-Dco / gDco) - Depending on the mode in which the invention is carried out, the yield can in particular be expressed for acids organic / VFAs having two to five carbon atoms or organic acids / AGVs having two to six carbon atoms.

By "selectivity" is meant the orientation towards the production of specific organic acids having from two to at least five carbon atoms. It is defined by the mass concentration ratio (g / g) of specific organic acids produced on the mass concentration of organic acids having from two to at least five carbon atoms produced. Depending on the case, the selectivity can be expressed as a function of a single specific organic acid or of several specific organic acids. The selectivity may especially be expressed as a function of specific organic acids being organic acids / AGVs having from four to five carbon atoms or organic acids / AGVs having from four to six carbon atoms.

The fermentable substrate can generally include any type of organic ways. The fermentable substrate is advantageously constituted by a or several effluents to be valorized and may, depending on the needs, be diluted or supplemented by nutrients necessary for the growth and maintenance of the microbial consortium, in particular nitrogen or phosphorus sources or any other nutrient identified as being present in limiting proportions. Preferably, the fermentable substrate is an effluent from the food industry, such as the dairy industry, sugar, starch and wheat.

The microbial consortium is a biomass operating under anaerobic and / or anoxic conditions. Biomass can come from a municipal digester, an industrial digester processing agro-food effluents or granular sludge from anaerobic reactor or rumens.

The single-stage reactor may be any type of reactor usually used in anaerobic / anoxic treatment. The method according to the invention is preferably implemented in continuous or semi-continuous mode in order to optimize productivity. The type of reactor is suitably selected depending on the type of fermentable substrate to be converted. The single-stage reactor may notably be chosen from the following list: a perfectly stirred reactor ("continuous stirred tank reactor" or "CSTR"), an upstream anaerobic sludge blanket reactor ("upflow anaerobic sludge blanket reactor" or "UASB"), an expanded granular sludge blanket (EGSB) anaerobic reactor, an anaerobic membrane reactor ("AnMBR") or a batch sequential reactor (" sequencing batch reactor "or" SBR ").

According to a particular embodiment of the invention, the advanced driving step comprises an initial phase of adaptation of the microbial consortium. This initial adaptation phase can last from a few hours to a few days, in particular from 0.5 days to 3 days. It consists of a batch fermentation of the substrate in contact with the inoculum. Following this initial acclimation step, the fermentation continues in semi-continuous or continuous mode with the application of an organic load progressively increased from a low initial value to a value higher target. The initial organic charge can in particular have a volume load of organic matter of between 5 and 20 gDCO / L / day and tend towards a target value of volume load of organic matter of between 20 and 80 gDCO / L / day depending on the substrate used. and desired organic acids.

The advanced driving stage is implemented through a control system. The control system comprises a set of sensors for the on-line measurement of the environmental parameters of fermentation and a control-command algorithm of the process which makes it possible to regulate in real time the reference values of the volume load of organic matter, of the time. hydraulic residence time, residence time of particulate matter, temperature and pH.

Environmental parameters measured online include:

the pH at the inlet and at the outlet of the single-stage reactor in order to estimate the rate of acidification in the single-stage reactor: the control of the pH in the single-stage reactor makes it possible in particular to avoid the inhibition of the production organic acids and to select the molecules to be produced;

temperature: the temperature control ensures that the fermentation is maintained at a mesophilic regime (20 - 40 ° C);

the mass flow rates of feed (s) and emptying (s) of the single-stage reactor: in particular, the mass feed rate of fermentable substrate makes it possible to estimate and / or measure the concentration and the flow of organic matter entering the fermenter.

The adjustment of the pH value can be controlled by adding an acid solution (for example HCl) and / or base (for example NaOH) and / or by controlling the level of alkalinity by addition of calcium carbonate. (CaCO ₃ ) or sodium carbonate (NaHCO ₃ ) for example.

The temperature adjustment can be controlled by heating means or cooling means.

The adjustment of the hydraulic residence time and / or the residence time of the particulate matter and / or the organic material volume load is controlled by the mass flow and / or volume flow (s) and discharge (s) of the single-stage reactor. Other measured parameters can be integrated into the control system in order to fine-tune the values of the organic matter load, the hydraulic residence time, the residence time of the particulate matter, the temperature and pH, with the ultimate aim of better controlling the development of microbial biomass and the nature of the organic acids produced. These other parameters can in particular be:

the electrical conductivity at the inlet and / or outlet of the single-stage reactor: this parameter may be correlated with the activity of the ions in the single-stage reactor;

the flow and composition of the gaseous effluent, in particular C0 ₂ , H ₂ , CH ₄ : a strong presence of methane can be correlated with an activation of archaea methanogens, which is not desired in the present invention; the presence of dihydrogen is also a good indicator of the progress of the fermentation stage and the nature of the organic acids produced (in particular the butyric acid / acetic acid ratio and the lactic acid concentration);

the partial pressure of C0 ₂ dissolves: the measurement of the partial pressure of dissolved carbon dioxide can serve as an indicator for the control of the alkalinity in the fermenter;

the oxidation-reduction potential: the oxidation-reduction potential is a good indicator of the anaerobiosis and the presence of oxygen in the single-stage reactor (microaeration conditions); it is also a good indicator of the quality of the organic acids produced; and,

the composition of organic acids, and in particular of AGVs, by automated titration, in-line (ion) chromatography and / or spectrometry; beforehand, the effluent at the outlet of the single-stage reactor must be filtered through a membrane to prevent clogging of the analyzer and reduce the time of maintenance ; automated titration analysis provides partial / total alkalinity and bicarbonate measurements.

In addition, to ensure anoxic / anaerobic conditions, nitrogen and / or carbon dioxide can be added in a controlled manner.

Since the industrial use of an on-line organic acid analyzer is currently only at the research stage, the inventors of the present invention have also developed a method for estimating organic acid composition from a set of measured parameters. More specifically, preferably, in the process according to the invention, the advanced step of conducting the acidogenic fermentation is modulated according to a continuous and real-time estimation of the concentration of organic acids in the single-stage reactor. from a set of measured parameters. Advantageously, the continuous estimation of the concentration of organic acids is carried out by partial least squares regression from on-line measurements of: the electrical conductivity, the partial pressure of dihydrogen and the flow rate of the gaseous effluent.

Preferably, the control-command system comprises at least one fuzzy logic algorithm. Fuzzy reasoning makes it possible to better understand natural phenomena that are imprecise and difficult to model by relying on the definition of rules and functions of belonging to so-called "fuzzy sets". It is for its ease of implementation and robustness that the fuzzy logic based regulation can be adopted for the control of the mixed culture anaerobic fermentation process developed in the present invention. Indeed, this method makes it possible to obtain a control law, often effective, without having to resort to important theoretical developments. It also presents the interest of taking into account the experiences acquired by the users and operators of the process to be controlled. The structure of a fuzzy controller is classically divided into three parts: fuzzification, the decision-making or inference mechanism, and defuzzification. Fuzzification is the stage that transforms a size measured on the process into a fuzzy set. The decision mechanism makes it possible to calculate the fuzzy sets associated with the command. Defuzzification is the step that transforms the fuzzy set, obtained by the previous calculation, into a control quantity to be applied to the process.

According to one particular embodiment of the invention, the organic acids of the mixture of organic acids have from two to six carbon atoms and the concentration of these organic acids having from two to six carbon atoms in the fluid effluent is strictly greater than 15 g / L.

Advantageously, the mass ratio of volatile fatty acids having from four to six carbon atoms relative to volatile fatty acids having from two to six carbon atoms is greater than or equal to 50% (weight, g / g). These concentration values of organic acids having from two to six carbon atoms and a mass ratio of volatile fatty acids having from four to six carbon atoms relative to fatty acids having from two to six carbon atoms can in particular be obtained by means of a method incorporating the method for estimating the organic acid composition from a set of measured parameters and the use of a fuzzy logic algorithm for the control system.

Thus, unlike the processes of the prior art, the process according to the present invention makes it possible to obtain an effluent having a high concentration of organic acids comprising organic acids with a high added value. In addition, the productivity of the process is advantageously greater than or equal to 7 g / L / day. The conversion efficiency of the process is advantageously greater than or equal to 0.3. Advantageously, the fermentable substrate and the microbial consortium present in the fluid effluent are recirculated, at least in part, in the single-stage reactor after solid / liquid separation of the organic acids having from two to at least five carbon atoms. of the liquid effluent. The solid / liquid separation may in particular be implemented by a membrane separator, a centrifuge or by another decanting means such as that included in the single-stage reactor (SBR, UASB, EGSB ...). This embodiment notably allows a decoupling between the HRT and the SRT.

The non-recirculated liquid fraction in the single-stage reactor can be sent to a step of recovery and purification of organic acids.

The present invention also relates to an installation comprising a single-stage reactor and a control-command system, comprising a set of sensors and at least one algorithm. The installation makes it possible in particular to implement the method described above.

5. List of figures

The invention, as well as the various advantages that it presents, will be more easily understood thanks to the following description of nonlimiting embodiments thereof, given with reference to the drawings, in which:

Figure 1 schematically shows an installation comprising a single-stage reactor CSTR for the implementation of the present invention.

FIGS. 2A and 2B show the composition of organic acids (volatile fatty acids and lactate), and ethanol for two tests carried out, with a variable organic charge (COD) and a regulation of the pH, the temperature and the hydraulic residence time .

Figure 3A is an estimate of the production of organic acids (volatile fatty acids and lactate) in a single-stage RI fermentation reactor that has evolved over time under different organic loading conditions (changes in feed concentration). . The variation of the conditions of the organic charge is shown in FIG. 3B.

Figure 4A is an estimate of the production of organic acids (AGV or VFA + lactate) on a single-stage fermentation reactor R2. This single-stage reactor has evolved over time under different pH conditions, organic load, type of fermentation deposit (substrates rich in sugars and / or proteins) and the addition or not of an external gas (H ₂ / C0 ₂ ). The variation of the conditions of the organic charge is shown in FIG. 4B.

Figure 5A is an estimate of the production of organic acids (AGV or VFA + lactate) on a single-stage R4 fermentation reactor. This single-stage reactor has evolved over time under different pH, organic load and reagent conditions used for pH control (NaOH, Na ₂ CO ₃ , CaCO ₃ and NH ₄ HCO ₃ ). The variation of the organic loading conditions is shown in FIG. 5B.

Figures 6A, 6B, 6C, 6D show performance results of a method using a fuzzy logic algorithm. In particular, FIG. 6A represents the evolution of the electrical conductivity (measurement-setpoint) and of the feed rate, FIG. 6B represents the evolution of the flow rate in dihydrogen (measurement-setpoint) and of the feed rate, the Figure 6C shows the evolution of the concentration of molasses in COD in the diet and Figure 6D the production of total AGVs, acetate (AA), butyrate (AB) lactate and ethanol.

6. Description of detailed embodiments of the invention

6.1 - Installation with a single-stage fermentation reactor

CSTR

With reference to FIG. 1, an installation 1 in which the process according to the invention can be implemented comprises a mixing / dilution unit 20 in which a fermentable substrate 200, to be upgraded, is mixed with water 210. The fermentable substrate 200 may especially be in the form of a mixture of a main substrate with one or more secondary substrates. The water 210 used in the mixing / dilution unit 20 may be water resulting from a process producing at a certain stage of the water or coming from the network. The mixture 220 thus obtained is then sent into a feed tank 30, and is then introduced into a single-stage reactor 10 at the fluid feed point 11 via a low pipe 32. homogenization of the substrate in the feed tank 30 by recirculation of the fluid in this same tank feeding 30 (duck). Alternatively, the feed tank 30 could be mixed and homogonized by means of a mechanical stirrer.

The single-stage reactor 10 is a perfectly stirred tank reactor (CSTR). It comprises a stirring means 9. It also comprises a fluid feed point 11, a feed point for gas 12, a fluid discharge point 13 and a gas discharge point 14. It comprises also a feed point 60 in acid 65 and a feed point 70 in base 75. It also comprises a purge for fluid 15 and a purge for gas 16. A bacterial inoculum is introduced into the single-stage reactor 10 at startup. of the process.

The single-stage reactor 10 is supplied continuously or semi-continuously by the mixture 220 at the fluid feed point 11. Gas (C0 ₂ / N ₂ ) can be introduced at the feed point for gas 12 to meet anaerobic / anoxic fermentation conditions. The pH and alkalinity can be adjusted during fermentation by addition of 65 (HCl) acid and / or 75 bases (NaOH and or CaCO ₃ and / or NaHCO ₃ ). The fluid effluent 300 is recovered continuously or semi-continuously at the fluid discharge point 13 and is then separated by a solids separation unit 40 into a first phase 310 comprising biomass and a residual organic charge and in a second phase 320 comprising organic acids consisting of two to six carbon atoms. These two phases 310, 320 may be recirculated, at least in part, in the single-stage reactor 10 at recirculation points 18, 19. The recirculation of the second phase 320 is used only in case of malfunctioning of the extraction and / or the recovery of organic acids.

The installation 1 also comprises a control system 100 of the single-stage reactor 10 and a control system 110 of the feed tank 30.

The control system 100 regulates the hydraulic residence time, the residence time in particulate matter, the temperature and the pH. he can also be used to regulate the stirring rate in the single-stage reactor 10, the alkalinity and the flow rates of incoming and outgoing gas.

The control-command system 110 makes it possible to regulate the volume load of organic matter as well as the level in the feed tank 30. It also makes it possible to control the dilution and the mixture of the substrates to be fermented.

The control-command systems 100, 110 comprise a set of sensors distributed over the entire installation 1 and in particular:

a sensor 499 for measuring the total organic carbon concentration (BioTector)

- 500 meters;

level sensors 501;

sensors of pH / temperature 502 and of oxidation-reduction potential 505;

an electrical conductivity / temperature sensor 503;

sensors 504 for measuring the total pressure in the gas phase of the single-stage reactor 10, the concentration of hydrogen in the gas phase, the concentration in

CH ₄ in the gas phase of the single-stage reactor 10, the CO ₂ concentration in the gas phase of the single-stage reactor 10; and,

a sensor 506 for measuring the C0 ₂ partial pressure dissolved at the outlet of the single-stage reactor 10.

Installation 1 also includes a set of pumps and valves (shown but not numbered in Figure 1).

6.2 - Experimental results with different fermentable substrates

The principles of the invention described above are now presented through several examples of experimental tests carried out with different fermentable substrates and under different conditions in the installation presented above.

The different fermentable substrates tested were:

· Substrate 1: very high in sugars (> 60% of total COD); • Substrate 2: high in protein (> 50% of total COD);

• Substrate 3: mixed composition: sugars + proteins

• Substrate 4: mixed composition: sugars + proteins

• Substrate 5: mixed composition: Mixture substrates 1 and 3

For all the experimental test examples presented, the resources used were diluted in order to impose the specific operating conditions concerned. The degree of dilution varies according to the initial concentration of the available resource (COD). For all of the detailed examples, the aim has been to direct the production of organic acids to volatile fatty acids.

Table 1 summarizes the results obtained for the different resources considered:

Table 1 In all the examples described in Table No. 1, the resulting AGVs composition,% C4-> C6, is greater than 50% (g / g). Similarly, the concentration of organic molecules having from 2 to 6 carbon molecules is strictly greater than 15 g / l and the volumetric productivity obtained for all the examples is greater than 7 g C2-C6 / L / d. The presentation of the various examples illustrates the range of profiles in organic molecules having 2 to 6 carbon molecules that can be produced, while respecting the criteria of concentration, productivity and yield claimed in the present invention and which are not not achieved in a combined way, as is the case here, under similar operating conditions, in the references available in the state of the art. It should also be noted that some experimental tests have led to a high concentration of organic acids (> 35 g / L) and the inhibitory phenomena by toxicity of organic acids have not been observed. 6.3 - Search for optimal conditions for AGV production

In order to define the ranges of the operating parameters (such as the pH, the organic load and the hydraulic residence time) to be applied to the fermenter to achieve the desired production criteria, several combinations of these parameters have been studied. It was chosen to illustrate this exercise of research of the optimal conditions of production of AGVs through the two following tests, describing the fermentation of the molasses.

6.3-a Test No. 1 (see Example 2 of Table 1)

Figure 2A shows the concentration of AGVs, lactic acid and ethanol as a function of the organic charge applied in the CSTR reactor. The operating conditions are as follows: T = 30 ° C, pH = 6, HRT = 2 d, OLR = 10-90 g COD / L / d.

The load was increased at startup of the fermenter (between 0 and 35 days) gradually to reach 40 g COD / L / day in order to stabilize the production of AGVs and acclimate the biomass. Between 35 days and 366 days, once adapted biomass, the applied load was varied from 30 to 90 g COD / l / day ^r in order to investigate the impact of this parameter on the fermentation performance of the substrate 1. The results of these tests show that It has been possible to achieve AGV concentrations above 35 gAGV / L in a single-stage reactor, with a C4-C6 composition> 50% (mass) and a productivity close to 20 gAGV / L / d, thanks to to the combined application of the specific operating conditions. On certain operating phases (OLR> 50 g COD / L / day), important lactic acid productions are also observed. The control of the organic filler and its maintenance at a lower load of 50 g COD / L / d often reduced the production of this acid.

6.3-b Test n ° 2

Figure 2B illustrates the time evolution of the production of AGVs, lactic acid and ethanol on a fermentation test of the substrate 1, for a progressive increase of the organic load (by step of 10gDCO / L / day) and in the presence of pH disturbances (pH less than 5). The operating conditions in the CSTR reactor are as follows: HRT 2j, T = 30 ° C, pH = 5.8, OLR = 10-40 g COD / L / d.

This experimental sequence carried out at a hydraulic residence time of 2 days made it possible to obtain information on the impact of the pH and the organic charge applied on the fermentation and to obtain results in line with the present invention. Indeed, a pH less than or equal to 5 induces a lower concentration of products (AGV) than a pH of 6, as shown by the results presented in Figure 2B. In addition, at each pH change (voluntary or sustained), the concentration of AGVs decreases sharply in favor of a production of lactic acid (lactic fermentation) and ethanol. This element illustrates the possibility of orienting selectivity towards the production of lactic acid.

The results of this test suggest precise control of pH and addition of alkalinity as claimed in this invention. PH control is essential to ensure the stability and robustness of the process. It has therefore been proposed in preferred features of the invention to set up a redundant pH control. This redundant pH control is achieved by setting up minimum of two pH sensors. The information emitted by the pH sensors is processed simultaneously to define the appropriate actions to be applied to acid and base metering pumps. 6.4 Online Estimation of Volatile Fatty Acid and Lactic Acid Production by Partial Least Squares Regression

An online estimator of the concentration of organic acids produced in fermentation has been developed (volatile fatty acids + lactic acid). Several statistical learning models were thus tested to develop this module for online estimation and / or prediction of organic acid concentration. These models are based on artificial intelligence techniques. They use as input all or some of the information emitted by the sensors mentioned in section 6.1 and require low maintenance times and / or costs.

Figure 3A illustrates the result of an estimation of the organic acid concentration using a Partial Least Square Regression (PLSR) regression model and conductivity, hydrogen flow rate and of hydrogen concentration measured on an RI reactor operating in CSTR mode at an organic load (OLR) variable from 10 to 80 g COD / L / d (Figure 3B), a pH controlled around 6, a temperature of 30 ° C and a hydraulic residence time of 2 days. It is possible to note that in a global way, this model captures very well the production dynamics of volatile fatty acids and lactic acid.

The same PLSR model is used to estimate the production of organic acids on other reactors online. Figures 4A and 5A illustrate the results of this estimation for variable OLR loads (Figures 4B and 5B). These fermentors were operated continuously for almost a year in mixed culture and under different operating conditions. It can clearly be seen from FIGS. 4A and 5A the ability of the PLSR model and simple measurements of electrical conductivity, hydrogen flow rate and hydrogen concentration in the gas phase to correctly reproduce the impact: Of a variation of the alkalinity,

Of a variation of the organic load,

From a leaching of biomass,

A change in the quality of the substrate on the concentration of AGVs - From the injection of a gas on the production of AGVs.

However, in some phases of the R2 reactor tests, the model underestimates the high concentrations of organic acids that are mainly related to high lactic acid production (between 50 to 74 g / 1). The addition of information on the sodium flow rate used for the pH control, the evolution of the redox potential and the hydraulic residence time is envisaged to improve the prediction of this model on certain operating phases.

6.5 Examples of process control algorithms applied to the process

The main characteristics of fuzzy logic controllers developed and prototyped in simulation and validated at laboratory scale are reported in the

Table 2 presented below:

Test and

Type of Variables of Variables of

N ° validation controller measures command

experimental

Conductivity,

Direct

Concentration H ₂ Flow

1 Fuzzy Logic

(gas) supply Yes

Control

Gas flow

Direct

Conductivity Flow

2 Fuzzy Logic Yes

power supply

control

Estimation Conductivity

AGV and Concentration H ₂ Flow

3 No

lactate (gas) feed

Fuzzy Logic Biogas flow Control

PH

Fuzzy Logic Acid Flow

4 Consumption of No

basic rate control

based

Table 2

Control-command strategy No. 1 is explained in detail below. It has been tested to evaluate its capacity to control in closed loop the flow of substrate 1 feeding an infinitely mixed fermentation reactor according to:

1- Measuring hydrogen concentration in the gas phase;

2- Measurement of the biogas flow;

3- Measurement of the electrical conductivity in the liquid phase.

The decision mechanism of this regulator is defined based on experimental test results conducted to study the fermentation of substrate 1 and mixtures (substrate 5) within the scope of the invention. An exemplary performance of the control law is shown in Figures 6A, 6B, 6C, 6D. The result presented corresponds to 80 days of continuous operation of a perfectly stirred reactor fed by the substrate 1. Various disturbances were considered in order to evaluate the ability of this control-command to maintain the target performance of the process. For this purpose, several dihydrogen flow instructions (Q. _H 2) (Figure 6B) and electrical conductivity (C _mS / _cm) (Figure 6A) were selected. These reference values were chosen on the basis of the knowledge of the process in order to produce organic acids with high added value (ie butyric acid for a fermentation of substrate 1), without destabilizing the performance of the process (ie minimizing the high concentrations of lactic acid and acetic acid) generally signifying inhibition and / or slowing down of AGV production. This first fuzzy logic controller was also subjected to disturbances in the concentrations of flux entering the reactor (FIG. 6C): that is to say concentrations of molasses in the feed. Between 0 and 20 days, the fuzzy controller is operated at a hydrogen flow rate of 20 L _H2 / d, while the reference value of electrical conductivity is varied from 33 to 29 mS / cm. Despite the high concentrations of COD in the reactor feed RI, there was no inhibition or instability of performance in terms of production of AGVs (Figure 6D). In addition, with the activation of this fuzzy regulation system, the production of lactic acid and acetic acid has been considerably reduced. The drop in the concentration of its two acids was accompanied by an increase in the production of butyric acid, which represents almost 95% of the total production of organic acids.

The high selectivity for butyric acid is related to the fuzzy rules put in place in the regulator's reasoning that are based on thresholds of electrical conductivity, hydrogen concentration and biogas flow. Indeed, the value of these parameters can directly inform the operator on the quality of the AGVs and a possible production of lactic acid.

Between 20 and 30 days, a sharp decline in the production of butyric acid is observed on the reactor. It is linked to a sharp drop in pH to reach a value of 4.5 due to a regulation fault during a weekend. The presence of the controller allowed a very fast restart of the production of butyric acid on this reactor, in particular by extending the hydraulic residence time for a better acclimatization of the biomass. Compared to other reactors that have experienced the same pH disturbances, the presence of this fuzzy logic controller has allowed a 50% time saving on this restart phase.

Between 20 and 60 days, the drop in the COD concentration fed to this reactor had a rather negative effect on the concentration of AGVs (a drop of about 10 g / 1). This phase of the tests clearly shows that an action on the feed rate (ie modulation of the organic load to be applied to the reactor by an action on the hydraulic residence time) is not sufficient to compensate for the impacts of a strong lower concentration of the substrate to be degraded. However, despite this drop in performance, the criteria of concentration, efficiency and productivity remain met. In conclusion, these experimental results illustrate the ability of the fuzzy logic controller to compensate for pH and organic charge disturbances. This controller is also able to maintain the desired performance of the process according to the invention and to guarantee a high selectivity of the biomass for butyric acid. Indeed, it was possible to maintain the system around a set point in concentration of butyric acid established without measuring its concentration directly. The proposed controller also made it possible to integrate the following expert knowledge into rules:

· Minimum guidelines for hydrogen flow to avoid low production of butyric acid

• The maximum limits of electrical conductivity to avoid the production of lactic acid

• Minimum hydraulic residence time to avoid leaching of biomass An extension of the regulation to add the information emitted by the oxidation-reduction potential is envisaged in the future for a finer control of the metabolic pathways (types of acids). organic products). Indeed, first results not illustrated in this patent application clearly show the possibility of controlling the metabolic pathways expressed by acting on this parameter by an external injection of a gas. To avoid a drop in yield, controlling the co-fermentation of two and / or more substrates with different concentrations is also envisaged as a means of offsetting the decrease in the amount of organic matter to be recovered.

The first results of the tests carried out with deposit mixtures (Substrate 5) of Table 2 illustrate this principle of control. The control of this mixture also allows a qualitative control of the volatile fatty acids to be produced.

Claims

A method for the acidogenic fermentation of a fermentable substrate by a microbial consortium in a single-stage reactor, said method being for providing at least one fluid effluent comprising a mixture of organic acids having from two to at least five carbon atoms and a gaseous effluent, characterized in that: said process comprises an advanced step of conducting the acidogenic fermentation implemented by means of a control-command system allowing the maintenance during said advanced driving step in said single reactor; storied:

- the volume load of organic matter to a value between 10 and

80 gDCO / Liter / day;

- the hydraulic residence time (HRT) to a value between 0.5 days and

4 days ;

- the particulate matter residence time (SRT) to a value between 0.5 days and 7 days; and

- the temperature to a value between 20 ° C and 40 ° C and,

- pH at a value between 5 and 7.

2. Acidogenic fermentation process according to claim 1, characterized in that said control system allows the maintenance during said advanced driving step in said single-stage reactor:

- the volume load of organic matter to a value between 30 and

6 gDCO / Liter / day;

- the hydraulic residence time to a value between 1 day and 3 days; - the residence time of the particulate matter to a value between 1 day and 4 days

- the temperature to a value between 30 ° C and 37 ° C and,

- pH at a value between 5.5 and 6.5.

3. Acidogenic fermentation process according to any one of the claims preceding, characterized in that the organic acids of said mixture of organic acids have from two to five carbon atoms; and, preferably, the organic acids of said mixture of organic acids have from two to six carbon atoms.

4. Acidogenic fermentation process according to any one of the preceding claims, characterized in that said advanced driving step comprises a phase of adaptation of said microbial consortium.

5. Acidogenic fermentation process according to any one of the preceding claims, characterized in that said single-stage reactor is used in continuous or semi-continuous mode.

6. Acidogenic fermentation process according to any one of the preceding claims, characterized in that said step of conducting advanced acidogenic fermentation is modulated according to a continuous and real-time estimation of the concentration of organic acids in said reactor single-stage from a set of measured parameters.

7. Acidogenic fermentation process according to claim 6, characterized in that said continuous estimation of the organic acid concentration is carried out by partial least squares regression, said set of parameters comprising on-line measurements of the electrical conductivity and online measurements of the dihydrogen partial pressure and the flow rate of said gaseous effluent.

8. Acidogenic fermentation process according to any one of the preceding claims, characterized in that said control system comprises at least one fuzzy logic algorithm.

9. Acidogenic fermentation process according to any one of the claims preceding, characterized in that said organic acids of said mixture of organic acids have from two to six atoms and,

the concentration of said organic acids having from two to six carbon atoms in said fluid effluent is strictly greater than 15 g / l.

10. Acidogenic fermentation process according to claim 9, characterized in that the mass ratio of volatile fatty acids having from four to six carbon atoms relative to said volatile fatty acids having from two to six carbon atoms is greater than or equal to 50% (mass, g / g).

11. Acidogenic fermentation process according to any one of the preceding claims, characterized in that said fermentable substrate and said microbial consortium present in said fluid effluent are recirculated, at least in part, in said reactor after a separation of said organic acids having two to at least five carbon atoms of said fluid effluent.

12. Installation (1) for carrying out a method according to any one of claims 1 to 11 comprising:

a single-stage reactor (10) and,

a control system (100, 110) comprising a set of sensors (499, 500, 501, 502, 503, 504, 505, 506) and at least one algorithm.