Classifications

• • C—CHEMISTRY; METALLURGY
  • C12—BIOCHEMISTRY; BEER; SPIRITS; WINE; VINEGAR; MICROBIOLOGY; ENZYMOLOGY; MUTATION OR GENETIC ENGINEERING
  • C12M—APPARATUS FOR ENZYMOLOGY OR MICROBIOLOGY; APPARATUS FOR CULTURING MICROORGANISMS FOR PRODUCING BIOMASS, FOR GROWING CELLS OR FOR OBTAINING FERMENTATION OR METABOLIC PRODUCTS, i.e. BIOREACTORS OR FERMENTERS
  • C12M41/00—Means for regulation, monitoring, measurement or control, e.g. flow regulation
  • C12M41/48—Automatic or computerized control

Definitions

• the invention relates to the field of digesters, also called biogas reactor or methanizer.
• the invention relates to a method and a device for predicting an indicator for monitoring the state of a digester.
• a digester aims to produce biogas by anaerobic digestion.
• Anaerobic digestion corresponds to a cascade of biochemical reactions that convert a fraction of organic matter into biogas, the remaining matter being called digestate.
• each degradation step is carried by distinct groups of micro-organisms, each having their own kinetics, optimal growth conditions, specific inhibitory compounds;
• the typical response time of a digester treating sludge and/or waste following, for example, a change in feed is of the order of several weeks.
• black box models based on the monitoring parameters of the digestion process.
• a “black box” model is known as a model type that does not take into account the phenomena internal to the system studied, here the digester.
• Models are then trained in order to link these bio-indicators with behaviors observed in the digesters.
• this type of model generally includes a very large number of parameters, such as parameters of biological kinetics, inhibition constants, yields, which must be adjusted using real data.
• parameters of biological kinetics, inhibition constants, yields which must be adjusted using real data.
• the model parameters are adjusted from existing data, for example monitoring data from the last months of a digester, by minimizing the difference between the model outputs, such as the biogas flow rate, and the actual measurements, and
• the model can then be used to predict the future behavior of the digester, for example by changing the mode or the composition of the feed.
• hybrid models adjusted to real data.
• the latter can be coupled with empirical models of the “black box” type.
• Another approach consists in using data external to the process to inject them into the model.
• the publication Charnier, C, Latrille, E., Jimenez, J., Torrijos, M., Sousbie, P., Miroux, J., Steyer, J. -P., 2017. Fast ADM1 implementation for the optimization of feeding strategy using near infrared spectroscopy.
• Water Research 122, 27-35 discloses a method during which we trained a partial least squares approach model, known as PLS, which can predict hydrolysis kinetics of a substrate from its near infrared spectrum.
• a computer-implemented prediction method of the evolution of a monitoring indicator of a plurality of monitoring indicators of the state of a digester adapted to produce a digestate and biogas comprising:
• mechanistic models make it possible to inject knowledge based on explicit hypotheses, such as compliance with the material balance, and ensure to have a result respecting the laws of physics.
• said machine learning module comprises a neural network or a decision tree or decision tree forest module or a partial least squares approach module or a support vector regression module.
• the adjustment parameters of the mechanistic model are constrained within a range of predetermined values or constrained according to a predetermined distribution of the adjustment parameters.
• the mechanistic model is not trained with aberrant or obviously erroneous tuning parameter values.
• said indicator for monitoring the state of said digester comprises a flow rate of biogas produced at the outlet of the digester.
• said automatic learning module comprises a neural network or a decision tree or forest of decision trees module or a partial least squares approach module or a vector-based regression module. support.
• a robust and efficient module is implemented to determine the adjustment parameters of the mechanistic model.
• said parameters relating to the input comprise a methanogenic potential parameter and/or parameters of splitting. These parameters indeed constitute relevant indicators on the quality of the input.
• said digester control parameters comprise a temperature of said digester and/or a hydraulic residence time and/or an organic load and/or a mode of feeding said digester. These parameters indeed constitute parameters that can be directly adjusted by the operators and/or designers of the digester.
• the invention also relates to a training method implemented by computer of an automatic learning module for a prediction method as described above, comprising:
• a calibration step of a mechanistic model in which the values or distributions of the adjustment parameters of said mechanistic model are determined according to said experimental data
• the invention also relates to a method for regulating the control parameters of a digester comprising:
• the invention also relates to a device for training an automatic learning module for predicting an indicator for monitoring the state of a digester adapted to produce a digestate and biogas, containing methane, from a input comprising:
• each set of experimental data being associated with a digester of the plurality of different digesters; each set of experimental data comprising values of parameters relating to a digester, values of parameters relating to a type of input and a value of said indicator for monitoring the state of the associated digester as a function of these parameters;
• Means for training at least one automatic learning module as a function of said plurality of sets of experimental data said automatic learning module being trained to determine, for each set of experimental data, said adjustment parameters said mechanistic model describing the kinetics of at least one of the acidogenesis, acetogenesis and methanogenesis stages or their combination as a function of the control parameters of said associated digester and of the parameters relating to the quality of the input and independently of the values of the indicators monitoring of said associated digester.
• the invention also relates to a device for implementing the prediction of an automatic learning module driven by a device as described above, to predict the evolution of the state of a digester adapted to produce a digestate and biogas from an input comprising:
• the invention also relates to a device for controlling a digester adapted to regulate parameters of said digester and/or parameters for regulating the feed mode of the digester according to the prediction obtained by said prediction method described above.
• FIG. 1 is a flowchart of the prediction method according to the invention
• FIG. 2 is a flowchart of a training method according to the invention.
• FIG. 3 is a schematic view of a device for implementing a prediction method according to the invention.
• FIG. 4 is a schematic view of a device for implementing a drive method according to the invention.
• FIG. 5 is a schematic view of a process for regulating the control parameters of a digester.
• the prediction method 1 aims to predict the evolution of the state of a digester, supplied with an input such as sludge from a purification station and producing digestate and biogas as output.
• the prediction is thus implemented for a particular digester 50, here called the digester 50 studied.
• the evolution of the state of a digester 50 also presents a strong inertia and a slow evolution over time, which causes particular difficulties in anticipating the deterioration of its performance in order to modify its input supply.
• the method includes a first step of defining a mechanistic model making it possible to predict the evolution of an indicator for monitoring the state of the digester.
• the definition of the mechanistic model can be a choice made by the user implementing the prediction method 1. However, as will be seen later, this choice must in fact have been made before the training 2 of the module of automatic learning, because the training is dependent on the mechanistic model chosen.
• the flow rate of biogas produced at the outlet of the digester 50 is the condition monitoring indicator that we seek to predict, because it constitutes a important indicator in terms of profitability of the digester.
• the invention is not limited to this particular type of indicator, which may also correspond, without limitation, to the concentration of volatile fatty acids in the digestate, or else the NH3 concentration of the digestate.
• the scientific literature includes many mechanistic models adapted to anaerobic digesters.
• Anaerobic Digestion Model No. 1 (ADM1) and its variants are known.
• ADM1 Anaerobic Digestion Model No. 1
• Such models make it possible in particular to simulate the anaerobic reactions of a digester 50 to predict the state of the digester, its performance or even its stability.
• the invention is not limited to any particular mechanistic model. Indeed, as will be explained later, the invention will involve an automatic learning module to determine the optimal adjustment parameters of the mechanistic model in order to be able to best simulate the internal reactions of the digester 50 studied.
• any mechanistic model adjustable by a plurality of predefined parameters can be implemented.
• the values supplied to the mechanistic model are generally time series, respectively of the control of the digester 50 and of the quality of the input.
• quality parameters of the input of its methanogenic potential, which is a relatively important criterion, and of its chemical oxygen demand.
• the quality of the input can also include its parameters of fractionation into sugars, proteins, lipids, metallic elements, elemental composition, etc.
• the control parameters of the digester 50 can include, by way of non-limiting examples, the temperature of the digester, the hydraulic residence time, the organic load, the mode of feeding of said digester, etc.
• a step 12 of determining the adjustment parameters of said mechanistic model is implemented.
• the adjustment parameters of a mechanistic model for an anaerobic digester will be used to calibrate the representation of the kinetics of the stages of anaerobic decomposition, namely: the stages of hydrolysis, acidogenesis, acetogenesis and methanogenesis.
• Hydrolysis refers to the step during which high molecular weight organic molecules such as polysaccharides, proteins and lipids are converted into monomers (e.g. glucose, fatty acids, amino acids). This step is commonly described by first-order kinetics associated with a hydrolysis constant. In particular, reference can be made for the description of this hydrolysis kinetics to the publication Vavilin, VA., Fernandez, B., Palatsi, J., Flotats, X., 2008. Hydrolysis kinetics in anaerobic degradation of particulate organic material : An overview. Waste Management 28, 939-951. https://doi. org/10.1016/j. wasman.2007.03.028
• Acidogenesis corresponds to the stage during which the monomers resulting from the hydrolysis stage are transformed into organic acids, alcohols, FL and CO2.
• Kinetics of this step can for example be described according to Monod's empirical equation, characterized by a maximum growth rate and an affinity constant.
• the Monod equation is notably described in the publication Liu, Y., 2007. Overview of some theoretical approaches for derivation of the Monod equation. Appl Microbiol Biotechnol 73, 1241-1250. https://doi.org/10.1007/s00253-006-0717-7.
• the acetogenesis phase allows the transformation of the various organic compounds from the two previous phases into acetate, H2 and CO2.
• the kinetics of this phase can also be described by Monod kinetics corrected for the accumulation of reaction products such as EU.
• the Monod kinetics thus corrected is described in particular in the publication Patôn, M., Rodriguez, J., 2019. Integration of bioenergetics in the ADM 1 and its impact on model predictions. Water Sci Technol 80, 339-346. https://doi.org/10.2166/wst.2019.279.
• the methanogenesis step corresponds to the conversion of EL, CO2 and/or acetate into CEE.
• the kinetics of this step can be described by Monod kinetics corrected for inhibition related to toxic compounds such as ammoniacal nitrogen.
• Rea description of the kinetics of this methanogenesis step reference may be made in particular to the publication Capson-Tojo, G., Moscoviz, R., Astals, S., Robles, A., Steyer, J.-R, 2020. Unraveling the literature chaos around free ammonia inhibition in anaerobic digestion. Renewable and Sustainable Energy Reviews 117, 109487. https://doi.Org/10.1016/j.rser.2019.109487_.
• the adjustment parameters of the mechanistic model describing the kinetics of at least one of the acidogenesis, acetogenesis and methanogenesis steps or their combination are determined.
• the determination step 12 is implemented by implementing an automatic learning module trained to determine the adjustment parameters according to the control parameters of the digester 50 studied and the parameters relating to the quality of the input.
• This automatic learning module thus receives as input control parameters of the digester 50 studied and parameters relating to the quality of the input and determines as output parameters for adjusting the digester.
• the automatic learning module comprises in particular a plurality of artificial neural networks trained to each detect a tuning parameter of the mechanistic model.
• An alternative approach can implement an artificial neural network trained to output a single vector of settings parameters.
• the invention is not limited to the implementation of neural networks, other automatic learning methods prove to be just as relevant for this type of learning. Mention will in particular be made of statistical regression methods such as support vector machines, SVM, decision trees, forests of decision trees, partial least squares regression or any other automatic learning method, supervised or not.
• the adjustment parameters of the mechanistic model are then retrieved. Consequently, a calibration step 13 of the mechanistic model is implemented with these determined parameters. During this calibration step 13, however, we want to make sure not to adjust the mechanistic model with aberrant parameter values. Also, we constrain the search for parameter values within a predetermined distribution.
• This method comprises a first step 21 of acquisition of a plurality of sets of experimental data.
• Each set of experimental data comprises time series of values of control parameters of a digester, values of parameters relating to the quality of a type of input supplying this digester 50 and time series of values of the indicator of monitoring of the digester, here its methane flow according to these parameters.
• the adjustment parameters obtained during this calibration step 21 constitute, with the set of experimental data associated with it, a set of training data for the automatic learning module.
• a training step 23 of the learning module according to the training data constituted by the couples sets of experimental data/adjustment parameters of the determined mechanistic model.
• the learning model will be able to evaluate, as a function of control parameters of said digester 50 and of parameters relating to the quality of the input supplying said digester, the adjustment parameters of the mechanistic model chosen in an optimized manner.
• This prediction method 1 can be implemented within the framework of more general methods, such as a method of regulation 53 of the control parameters of a digester 50 represented in figure 5.
• Such a regulation method 53 comprises the acquisition of a prediction obtained by the implementation of the prediction method 1.
• the implementation device 40 of the prediction method 1 is preferably a computer, comprising a processor 42 for carrying out the calculations, a random access memory 41, a data storage medium 43, input-output communication means 46, these means being adapted to store in memory, either in RAM 41 or on the storage medium 43, a mechanistic model 48 and a machine learning module 45.
• the input-output communication means 46 can be of any known type, Ethernet port, COM port, USB port, wired or wireless communication device.
• the digester control device 50 is generally, but not necessarily, of the same type as the implementation device 40 of the prediction method 1.
• the training device 30 of the training method 2 is generally, but not necessarily, a separate device from the implementing device 40 of the prediction method 1.
• This training device 30 also comprises a computer, comprising a processor 42 for carrying out the calculations, a random access memory 41, a data storage medium 43, input-output communication means 46, these means being suitable for storing in memory , either in RAM 41 or on the storage medium 43, a mechanistic model 48 and a machine learning module 45.
• a computer comprising a processor 42 for carrying out the calculations, a random access memory 41, a data storage medium 43, input-output communication means 46, these means being suitable for storing in memory , either in RAM 41 or on the storage medium 43, a mechanistic model 48 and a machine learning module 45.
• the input-output communication means 46 can be of any known type, Ethernet port, COM port, USB port, wired or wireless communication device.
• the training device 30 can implement the training steps of the automatic learning module with parallel computing processors, such as GPUs implemented in a general computing context, known in English as acronym for GPGPU, for General-Purpose computing on Graphics Processing Units.
• parallel computing processors such as GPUs implemented in a general computing context, known in English as acronym for GPGPU, for General-Purpose computing on Graphics Processing Units.
• the adjustment of the mechanistic models 38, 48 can be carried out using Bayesian approaches in order to transform the deterministic mechanistic model into a stochastic model.

Landscapes

• Chemical & Material Sciences (AREA)
• Engineering & Computer Science (AREA)
• Analytical Chemistry (AREA)
• Organic Chemistry (AREA)
• Health & Medical Sciences (AREA)
• Life Sciences & Earth Sciences (AREA)
• Bioinformatics & Cheminformatics (AREA)
• Zoology (AREA)
• Wood Science & Technology (AREA)
• Sustainable Development (AREA)
• Microbiology (AREA)
• Biotechnology (AREA)
• Biomedical Technology (AREA)
• Biochemistry (AREA)
• General Engineering & Computer Science (AREA)
• General Health & Medical Sciences (AREA)
• Genetics & Genomics (AREA)
• Computer Hardware Design (AREA)
• Treatment Of Sludge (AREA)
• Feedback Control In General (AREA)

Applications Claiming Priority (2)

Publications (1)

Family

ID=80446788

Family Applications (1)

Country Status (4)

Family Cites Families (3)

• 2021
  • 2021-08-12 FR FR2108662A patent/FR3126127B1/fr active Active
• 2022
  • 2022-08-12 CN CN202280055984.6A patent/CN117813371A/zh active Pending
  • 2022-08-12 EP EP22768447.9A patent/EP4384596A1/de active Pending
  • 2022-08-12 WO PCT/FR2022/051580 patent/WO2023017229A1/fr not_active Ceased

Also Published As

Similar Documents

Legal Events