JP7142501B2 - Method for creating prediction information for predicting amount of biogas generation, and use of the prediction information - Google Patents

Description

TECHNICAL FIELD The present invention relates to a method of creating prediction information for predicting the amount of biogas generated, and use of the prediction information.

Fermentation phenomena caused by microorganisms are useful phenomena utilized in various fields. For example, fermentation phenomena are used in the fields of bioenergy production, wastewater treatment, and food production.

Producing a fermentation phenomenon in a desired state is very important from the viewpoints of efficiently producing bioenergy, efficiently treating wastewater, and efficiently producing food. In order to bring about a fermentation phenomenon in a desired state, it is first necessary to accurately grasp the state inside the fermenter. Therefore, conventionally, various techniques have been developed for grasping the state inside the fermenter.

For example, Patent Literature 1 describes an anaerobic treatment method for biodegrading an object to be treated containing cellulosic organic matter. In the anaerobic treatment method, the state inside the fermenter is grasped based on the number of cellulose-degrading bacteria, and the anaerobic treatment is controlled accordingly.

Patent No. 4354311 (published on August 7, 2009)

However, the conventional technology as described above has a problem that the state inside the fermenter cannot be accurately grasped.

Specifically, in the prior art, no method has been established for selecting important information from a large amount of information in the fermenter. Furthermore, in the prior art, even if important information can be selected, a method for evaluating the information has not been established. Therefore, the prior art has the problem that the conditions in the fermenter cannot be accurately grasped.

An object of one aspect of the present invention is to provide a method of creating prediction information for accurately grasping the state inside a fermenter, and use of the prediction information.

As a result of intensive studies in view of the above problems, the present inventors have found that the state in the fermenter can be accurately determined by using a function containing bacterial flora data on two or more types of bacteria and data on the amount of biogas generated as variables. The present inventors have found that it is possible to comprehend, and have completed the present invention.

In order to solve the above problems, a creation method according to one aspect of the present invention is a method for creating prediction information for predicting the amount of biogas generated, the method comprising: a measurement data acquisition step of acquiring a plurality of measurement data including data and bacterial flora data on two or more types of bacteria in the fermenter when the data on the amount of biogas generated is acquired; and the measurement data. a matrix data creation step of creating matrix data using at least two or more of the above, the bacterial flora data as an explanatory variable, and the biogas generation amount data as an objective variable; and an analysis process for the matrix data. and a matrix data processing step of obtaining, as the prediction information, a function containing bacterial flora data on the two or more types of bacteria and data on the amount of biogas generated as variables.

A function obtained from bacterial flora data on two or more types of bacteria contains a lot of information reflecting the state of the fermenter. By substituting the bacterial flora data obtained from the actually operated fermenter into the function, not only can the amount of biogas generated in the fermenter be accurately calculated, but also based on the amount of biogas generated , the conditions in the fermentation tank can be accurately grasped. If the state in the fermenter can be accurately grasped, the operation method of the fermenter can be changed as necessary, thereby adjusting the state in the fermenter to the desired state (for example, the target substance in the fermenter can be can be processed efficiently by

In the creation method according to one aspect of the present invention, the matrix data processing step performs multiple regression analysis on the matrix data, and obtains bacterial flora data on the two or more types of bacteria, and the amount of biogas generated. It is preferable that the step of obtaining a multiple regression equation including data as variables as the prediction information.

The multiple regression equation is a function that can highly accurately associate bacterial flora data on two or more types of bacteria with data on the amount of biogas generated. By substituting bacterial flora data obtained from an actually operated fermenter into the multiple regression equation, not only can the amount of biogas generated in the fermenter be calculated more accurately, but also the amount of biogas generated can be calculated. based on this, the conditions in the fermenter can be grasped more accurately.

In the creation method according to one aspect of the present invention, after the matrix data processing step, (i) the bacterial flora data A of the measurement data not included in the matrix data is substituted into the function, and the amount of biogas generated and (ii) a comparison step of comparing the calculated value with the measured value of the amount of biogas generated from the fermenter when the bacterial flora data A was acquired. It is preferable to have an accuracy check step included.

The calculated value of the amount of biogas generated by substituting the bacterial flora data A of the measurement data not included in the matrix data into the above function, and the measured value of the amount of biogas generated when the bacterial flora data A was obtained. The accuracy of the function can be known by comparing with (the measured value). That is, if the difference between the calculated value and the measured value is small, it can be determined that the accuracy of the function is high, and if the difference between the calculated value and the measured value is large, it can be determined that the accuracy of the function is low. If the precision of the function is known, the matrix data can be recreated as necessary to obtain a function with higher precision.

In the creation method according to one aspect of the present invention, the bacterial flora data includes microorganisms belonging to the phylum Euryarchaeota, microorganisms belonging to the phylum Bacteroidetes, microorganisms belonging to the phylum Firmicutes, and microorganisms belonging to the phylum OP9. It preferably includes microbiota data of at least one microorganism selected from the group consisting of microorganisms, microorganisms belonging to the phylum Synergistetes, and microorganisms belonging to the phylum Thermotogae.

The microorganisms described above are microorganisms that are greatly involved in the fermentation phenomenon that generates biogas. Therefore, with the above configuration, the state inside the fermenter can be grasped more accurately.

In order to solve the above problems, a processing method according to one aspect of the present invention is a processing method in which a target substance in a fermenter is treated with bacteria while generating biogas, A bacterial flora data acquisition step of acquiring bacterial flora data relating to two or more types of bacteria, the bacterial flora data relating to the two or more types of bacteria, which are prepared in advance, and the generation of biogas. and a calculating step of calculating a predicted generation amount of biogas by substituting the amount data into a function containing the amount data as a variable.

A function obtained from bacterial flora data on two or more types of bacteria contains a lot of information reflecting the state of the fermenter. By substituting the bacterial flora data obtained from the actually operated fermenter into the function, not only can the amount of biogas generated in the fermenter be accurately calculated, but also based on the amount of biogas generated , the conditions in the fermentation tank can be accurately grasped. If the state in the fermenter can be accurately grasped, the operation method of the fermenter can be changed as necessary, thereby adjusting the state in the fermenter to the desired state (for example, the target substance in the fermenter can be can be processed efficiently by

In the processing method according to one aspect of the present invention, in the calculation step, the bacterial flora data is prepared in advance, and the bacterial flora data related to the two or more types of bacteria and the biogas generation amount data are calculated. It is preferable that the process is a step of calculating the predicted generation amount of biogas by substituting into a multiple regression equation that is included as a variable.

The multiple regression equation is a function that can highly accurately associate bacterial flora data on two or more types of bacteria with data on the amount of biogas generated. By substituting bacterial flora data obtained from an actually operated fermenter into the multiple regression equation, not only can the amount of biogas generated in the fermenter be calculated more accurately, but also the amount of biogas generated can be calculated. based on this, the conditions in the fermenter can be grasped more accurately.

The processing method according to an aspect of the present invention preferably includes a state changing step of changing the state inside the fermenter after the calculating step.

According to the above configuration, the state in the fermenter can be brought closer to a desired state (for example, a state in which the target substance in the fermenter can be efficiently treated with bacteria).

In order to solve the above problems, a processing system according to one aspect of the present invention is a processing system that processes a target substance with bacteria while generating biogas, wherein the target substance and the bacteria are mixed. a fermenter for doing so, a bacterial lawn data acquisition unit for acquiring bacterial lawn data on two or more types of bacteria in the fermenter, and the bacterial lawn data, which are created in advance, the two or more types and a calculation unit for calculating a predicted biogas generation amount by substituting the bacteria flora data related to bacteria and the biogas generation amount data as variables into a function. there is

A function obtained from bacterial flora data on two or more types of bacteria contains a lot of information reflecting the state of the fermenter. By substituting the bacterial flora data obtained from the actually operated fermenter into the function, not only can the amount of biogas generated in the fermenter be accurately calculated, but also based on the amount of biogas generated , the conditions in the fermentation tank can be accurately grasped. If the state in the fermenter can be accurately grasped, the operation method of the fermenter can be changed as necessary, thereby adjusting the state in the fermenter to the desired state (for example, the target substance in the fermenter can be can be processed efficiently by

In the processing system according to the aspect of the present invention, the calculation unit converts the bacterial flora data into the bacterial flora data related to the two or more types of bacteria, which have been created in advance, and the biogas generation amount data. It is preferable to calculate the predicted generation amount of biogas by substituting into the multiple regression equation included as a variable.

The multiple regression equation is a function that can highly accurately associate bacterial flora data on two or more types of bacteria with data on the amount of biogas generated. By substituting bacterial flora data obtained from an actually operated fermenter into the multiple regression equation, not only can the amount of biogas generated in the fermenter be calculated more accurately, but also the amount of biogas generated can be calculated. based on this, the conditions in the fermenter can be more accurately grasped.

The processing system according to one aspect of the present invention preferably further includes a state changing section that changes the state inside the fermenter.

According to the above configuration, the state in the fermenter can be brought closer to a desired state (for example, a state in which the target substance in the fermenter can be efficiently treated with bacteria).

According to one aspect of the present invention, it is possible to provide prediction information that can accurately calculate the amount of biogas generated and accurately grasp the state in the fermenter based on the amount of biogas generated. .

According to one aspect of the present invention, by using the prediction information, the target substance in the fermenter is treated with the fungus in a desired state (for example, a state in which the target substance in the fermenter can be efficiently treated with the fungus). It is possible to provide a processing method that can

According to one aspect of the present invention, by using the prediction information, the target substance in the fermenter is treated with the fungus in a desired state (for example, a state in which the target substance in the fermenter can be efficiently treated with the fungus). It is possible to provide a processing system capable of

(A) and (B) are flowcharts showing the procedure of a prediction information creation method according to an embodiment of the present invention. (A) to (D) are flowcharts showing procedures of a processing method according to one embodiment of the present invention. It is a mimetic diagram showing composition of a processing system concerning one embodiment of the present invention. It is a mimetic diagram showing composition of a processing system concerning one embodiment of the present invention. It is a graph which shows the analysis result in the Example of this invention. It is a graph which shows the analysis result in the Example of this invention. It is a graph which shows the analysis result in the Example of this invention. It is a graph which shows the analysis result in the Example of this invention.

One embodiment of the present invention is described below, but the present invention is not limited thereto. The present invention is not limited to the configurations described below, and can be modified in various ways within the scope of the claims, and the technical means disclosed in different embodiments and examples. Embodiments and examples obtained by appropriate combinations are also included in the technical scope of the present invention. Also, all of the documents mentioned in this specification are incorporated herein by reference. In this specification, when "A to B" is described with respect to a numerical range, the description means "A or more and B or less".

[1. Prediction information creation method]
In the creation method of the present embodiment, prediction information (specifically, a function) for predicting the amount of biogas generated is created. The amount of biogas generated reflects the conditions in the fermenter. Therefore, by using the prediction information, it is possible to accurately grasp the state inside the fermenter (for example, it is possible to quickly detect the operating efficiency of the fermenter, which is declining). Furthermore, if the state inside the fermenter can be accurately grasped, the state inside the fermenter can be changed to a desired state. Furthermore, if the state in the fermenter can be changed to a desired state, for example, (i) the production efficiency of biogas is increased, (ii) raw materials (e.g., waste water) processed per unit time in the fermenter (iii) increase the amount of product (e.g., bioenergy such as ethanol, or food) produced per unit time in the fermenter, or (iv) the operating efficiency of the fermenter ( For example, it is possible to suppress a decrease in product production efficiency).

As used herein, the term "biogas" refers to gas generated by the action of organisms or substances derived from organisms (eg, enzymes). Examples of biogases include, but are not limited to, methane gas, carbon dioxide, ammonia, hydrogen sulfide, hydrogen, water vapor, and acetic acid. When biogas is methane gas and/or carbon dioxide, the state of methane fermentation in the fermenter can be accurately grasped. When the biogas contains hydrogen and/or acetic acid, the state of hydrogen fermentation and/or acetic acid fermentation in the fermenter can be accurately grasped. Therefore, the prediction information may be for predicting the amount of biogas generated during methane fermentation, hydrogen fermentation, and/or acetic acid fermentation.

The method of creating prediction information for predicting the amount of biogas generated according to the present embodiment comprises at least a measurement data acquisition step (S1), a matrix data creation step (S2), as shown in FIG. , and a matrix data processing step (S3), and each step is performed in this order. In addition, as shown in (B) of FIG. 1, the method of creating prediction information for predicting the amount of biogas generated according to the present embodiment includes, in addition to the above steps, an accuracy confirmation step (S4 ), and the accuracy confirmation step may be performed after the matrix data processing step. Each step will be described below.

<Measurement data acquisition process>
The measurement data acquisition step includes measurement data including data on the amount of biogas generated from the fermenter and bacterial flora data on two or more types of bacteria in the fermenter when the data on the amount of biogas generated was acquired. is a step of acquiring a plurality of

In this case, the measurement data include the pH value in the fermenter, the organic oxygen concentration in the fermenter (more specifically, data on the amount of organic oxygen generated), the ammonia concentration in the fermenter (more specifically, ammonia yield data), the temperature in the fermenter, and/or the amount of raw material introduced into the fermenter. According to this configuration, it is possible to more accurately grasp the state inside the fermenter.

The measurement data consists of data on the amount of biogas generated from the fermenter, and bacterial flora data on two or more types of bacteria in the fermenter when the data on the amount of biogas generated was acquired. may

In the measurement data acquisition step, a single piece of data on the amount of biogas generated may be treated as "data on the amount of biogas generated derived from the fermenter". In addition, in the measurement data acquisition process, each of the plurality of biogas generation amount data or the sum of the plurality of biogas generation amount data is treated as "data on the biogas generation amount derived from the fermentation tank". may For example, methane fermentation mainly produces methane gas and carbon dioxide. In this case, in the measurement data acquisition step, the amount of methane gas generated, the amount of carbon dioxide generated, or the sum of the amount of methane gas generated and the amount of carbon dioxide generated is "data of the amount of biogas generated from the fermentation tank". can be treated as For example, let V(L) be the generated amount of biogas containing n kinds of gases, and let R _n (%) be the abundance ratio of each gas in the biogas. At this time, the relational expression “V=V×(R ₁ /100)+V×(R ₂ /100)+ . . . +V×(R _n /100)” holds. In this case, "V x (R ₁ /100)", "V x (R ₂ /100)", ..., "V x (R _n /100)" or "V" It may be treated as "data on the amount of biogas generated". In addition, the total amount of generated two or more types of gas (for example, two types, three types, ..., or n-1 types) arbitrarily selected from n types of gas is defined as "in the fermenter It may be treated as "data of the generated amount of biogas derived from".

Data on the amount of biogas generated can be obtained by a known method. Devices are commercially available for measuring the production of specific gases. Therefore, data on the amount of biogas generated may be obtained using the device.

The amount of biogas generated may be the amount of biogas newly generated in the fermenter within a predetermined period of time around the time of measurement. In this case, the amount of biogas generated is, for example, 1 second, 5 seconds, 10 seconds, 30 seconds, 1 minute, 5 minutes, 10 minutes, 15 minutes, or 30 minutes. It may be the amount of gas, but the invention is not limited to this. Also, the amount of biogas generated may be the amount of biogas accumulated inside or outside the fermenter (for example, a storage tank) from the start of operation of the fermenter to the time of measurement. In this case, the amount of biogas generated is, for example, 1 hour, 5 hours, 10 hours, 1 day, 3 days, 5 days, or 10 days after the start of operation of the fermenter. It may be the amount of biogas accumulated, but the invention is not so limited.

As used herein, the term "bacteria flora data" means the ratio of specific microorganisms to all microorganisms (in other words, bacteria, etc.) contained in a culture sample. For example, when all microorganisms are quantified as 1, if microorganism A is quantified as 0.1 and microorganism B is quantified as 0.3, the bacterial flora data for microorganism A is 0.1, and the bacterial flora for microorganism B is 0.1. The data is 0.3. Further processing of these bacterial flora data may be used as "bacterial flora data" in the matrix data creation step and the matrix data processing step, which will be described later. The specific content of the processing is not particularly limited, and processing such as centering, scaling, logarithmic conversion, or square root can be performed. For example, it is preferable to perform this process when forming the measurement data using n−1 or less types of bacterial lawn data among the acquired n types of bacterial lawn data. According to this configuration, it is possible to more accurately grasp the state inside the fermenter. In addition, when describing "microorganism" and "bacteria" in this specification, unless otherwise defined, both descriptions intend "microorganism" broadly.

When acquiring bacterial flora data, bacteria may be identified and quantified based on the nucleotide sequence of a gene or the amino acid sequence of a protein. Examples of such genes (or proteins) include 5S rRNA, 5.8S rRNA, 12S rRNA, 15S rRNA, 16S rRNA, 18S rRNA, 23S rRNA, and 28S rRNA. It is not limited to this.

Bacterial flora data can be acquired by a known method. For example, (a) obtaining a culture sample from a fermenter, (b) obtaining nucleic acid (e.g., genomic DNA or RNA) from a bacterium contained in the culture sample, and (c) identifying the bacterium Amplifying (e.g., quantitatively amplifying) a region within the nucleic acid that is available for microbial flora, and (d) obtaining bacterial flora data using the amplified nucleic acid, through Data can be obtained.

The above step (a) can be performed, for example, by obtaining a culture sample in the fermenter through an opening provided in the fermenter. The above step (b) can be performed using a commercially available nucleic acid extraction kit. The above step (c) can be performed by a known method such as a PCR (Polymerase Chain Reaction) method. The step (d) above can be performed by known methods such as NGS (Next Generation Sequencer), DGGE (Denaturing Gradient Gel electrophoresis), RT-PCR (Reverse Transcription Polymerase Chain Reaction), and/or microarray. . In addition, the above step (d) may be performed using a commercially available kit. In addition, at the time of the filing of the present application, there are many companies that have commissioned the analysis of bacterial flora data. Therefore, bacterial flora data can be obtained by consigning the work (for example, the above-described steps (b) to (d)) to these companies. Since the above steps (a) to (d) can all be performed by known methods, detailed description thereof will be omitted in this specification.

The number of types of bacteria for which bacterial flora data is acquired may be two or more, for example, five or more, ten or more, twenty or more, thirty or more, forty or more, or even fifty or more. good. The upper limit of the types of bacteria for which bacterial flora data is acquired is not particularly limited, but may be, for example, 1000 types, 700 types, 500 types, 300 types, 200 types, 100 types, or 50 types.

When maintaining the state in the fermenter in a desired state (for example, a state in which the target substance in the fermenter can be efficiently processed by the fungus), not all the fungi are involved in maintaining the state in the fermenter. . Therefore, it is preferable that the bacteria for which bacterial flora data is acquired include bacteria that have a large contribution to maintaining the state in the fermenter (in other words, bacteria that have a large contribution to the fermentation phenomenon). It is more preferable that 1/10 or more of the bacteria for which flora data is obtained are bacteria that are greatly involved in maintaining the state in the fermenter, and 1/5 or more of the bacteria for which bacterial flora data is obtained are in the state in the fermenter. It is more preferable that the bacteria are greatly involved in the maintenance of the microbial flora, and it is further preferable that 1/2 or more of the bacteria for which the bacterial flora data is acquired are the bacteria that are greatly involved in maintaining the state in the fermenter. It is more preferable that 4/5 or more of the bacteria for which data is acquired are bacteria that are greatly involved in maintaining the state in the fermenter, and all the bacteria for which bacterial flora data is acquired are involved in maintaining the state in the fermenter. Most preferably, it is a bacterium that has a large involvement.

Microorganisms belonging to the phylum Euryarchaeota are the bacteria for which bacterial flora data are to be obtained, from the viewpoint that they are greatly involved in fermentation phenomena that generate biogas (e.g., methane fermentation, hydrogen fermentation, or acetic acid fermentation). , a microorganism belonging to the phylum Bacteroidetes, a microorganism belonging to the phylum Firmicutes, a microorganism belonging to the phylum OP9, a microorganism belonging to the phylum Synergistetes, and a microorganism belonging to the phylum Thermotogae. preferably contains at least one microorganism that

From the viewpoint that they are greatly involved in the fermentation phenomenon that generates biogas, more specifically, microorganisms for which bacterial flora data are obtained include microorganisms belonging to the phylum Euryarchaeota and the family Methanobacteriaceae; phylum Euryarchaeota microorganisms belonging to the genus Methanothermobacter; phylum Bacteroidetes microorganisms belonging to the order Bacteroidales; phylum Bacteroidetes microorganisms belonging to the family Porphyromonadaceae; Firmicutes ) phylum Microorganisms belonging to the genus Leuconostoc; phylum Firmicutes Microorganisms belonging to the order Clostridiales; phylum Firmicutes Microorganisms belonging to the family Clostridiaceae; phylum Firmicutes Clostridium Microorganisms belonging to the genus; Firmicutes phylum Microorganisms belonging to the Lachnospiraceae family; Firmicutes phylum Microorganisms belonging to the genus Pelotomaculum; Firmicutes Microorganisms belonging to the genus Syntrophomonas; Phylum Firmicutes (Sporomus) Microorganisms belonging to the genus Sporomusa; (Thermoanaerobacterales) microorganisms belonging to the order Firmicutes Microorganisms belonging to the genus Thermacetogenium; at least one microorganism selected from the group consisting of microorganisms belonging to the family erobaculaceae; microorganisms belonging to the phylum Synergistetes genus Anaerobaculum; and microorganisms belonging to the phylum Thermotogae genus S1 is preferred.

In the measurement data acquisition step, a plurality of pieces of measurement data including data on the amount of biogas generated and bacterial flora data on two or more types of bacteria are acquired. The number of measurement data may be 2 or more, and may be, for example, 5 or more, 10 or more, 20 or more, 30 or more, 40 or more, or 50 or more. The upper limit of the number of measurement data is not particularly limited, but may be 1000, 700, 500, 300, 200, 100, or 50, for example.

<Matrix data creation process>
The matrix data creation step is a step of creating matrix data using at least two or more of the measurement data, with bacterial flora data as an explanatory variable, and biogas generation amount data as an objective variable.

Although the form of the matrix data is not particularly limited, for example, the following matrix data can be mentioned. Of course, the form of matrix data of the present invention is not limited to the following examples.

The matrix data describes n measurement data. The number of measurement data (in other words, the value of n) may be 2 or more, for example, 5 or more, 10 or more, 20 or more, 30 or more, 40 or more, or 50 or more. may be The upper limit of the number of measurement data is not particularly limited, but may be 1000, 700, 500, 300, 200, 100, or 50, for example.

The matrix data describes the microbiota data “Xm” of m types of bacteria. The number of types of bacteria (in other words, the value of m) may be 2 or more, for example, 5 or more, 10 or more, 20 or more, 30 or more, 40 or more, or 50 or more. There may be. The upper limit of the types of bacteria for which bacterial flora data is acquired is not particularly limited, but may be, for example, 1000 types, 700 types, 500 types, 300 types, 200 types, 100 types, or 50 types.

The matrix data describes microbial flora data “Xm _n ” of each bacterium corresponding to n pieces of measurement data. The number of bacterial lawn data is determined according to the number of measurement data (in other words, the value of n) and the type of bacteria (in other words, the value of m) described above.

The matrix data describes the biogas generation amount data “Y _n ” corresponding to the n pieces of measurement data. The number of pieces of biogas generation amount data is determined according to the number of pieces of measurement data described above. The unit of the biogas generation amount data is not particularly limited, and may be, for example, "m ³ ", "L", "g", or "kg".

<Matrix data processing process>
The matrix data processing step is a step of performing an analysis process on the matrix data and obtaining, as the prediction information, a function containing bacterial flora data on two or more types of bacteria and data on the amount of biogas generated as variables.

The analysis process performed on the matrix data is not particularly limited as long as it is an analysis process by a machine learning method that can construct a prediction model. Examples of the machine learning method include multiple regression analysis (e.g., MVR, PCR, PLS, O-PLS, etc.), algorithms imitating human neural networks (e.g., perceptron, neural network, convolutional neural network, deep learning, etc.), logistic regression (e.g., coordinate descent method, gradient descent method, Newton method , and quasi-Newton methods, etc.), and support vector machines (eg, support vector regression, etc.).

In the matrix data processing step, multiple regression analysis is performed on the matrix data, and a multiple regression equation containing bacterial flora data on two or more types of bacteria and data on the amount of biogas generated as variables can be obtained as prediction information. can.

Specific examples of the method of multiple regression analysis are not particularly limited, and include MVR (Multivariate Regression), PCR (Principal Component Regression), PLS (Partial Least Squares), and O-PLS (Orthogonal PLS). can be done. From the viewpoint of obtaining a more accurate multiple regression equation, the method of multiple regression analysis is preferably MVR, more preferably PCR, more preferably PLS, and more preferably O-PLS. However, in the matrix data processing step, a highly accurate multiple regression equation can be obtained even with a multiple regression analysis method with low accuracy.

These multiple regression analysis methods are well known. For example, the matrix data processing step can be performed according to the method described in "Multivariate Analysis for Chemists (Introduction to Chemometrics) by Kodansha Scientific" written by Yukihiro Ozaki, Akifumi Uda and Toshio Akai. Said "Yukihiro Ozaki, Akifumi Uda, Toshio Akai, Multivariate Analysis for Chemists (Introduction to Chemometrics) Kodansha Scientific" is incorporated herein by reference. Multiple regression analysis can also be performed using known software. Examples of such software include Infometrix multivariate analysis software “pirouette series” and Umetrics multivariate analysis software “SIMCA series”. In this case, multiple regression analysis may be performed according to the protocol attached to the software for multivariate analysis. Also, the matrix data processing step may be performed using programming using the R language or Python language, or Matlab using the C language, but the present invention is not limited to these.

The multiple regression equation obtained by the multiple regression analysis is not particularly limited, and for example, the following multiple regression equation can be mentioned. In addition, the following multiple regression formula is an example, and the present invention is not limited to the multiple regression formula;
Y= _{b1X1 + b2X2} +...+ _bmXm +b0... (multiple regression equation);
In the above multiple regression equation, b ₁ to b _m and b ₀ indicate constants obtained _by multiple regression analysis, X1 to Xm indicate the bacterial flora data of m types of bacteria, and Y is biogas. It shows the data of the amount generated. By substituting the actually measured values of X1 to Xm into the multiple regression equation, the predicted value of the amount of biogas generated can be obtained. Then, based on the predicted value, the state inside the fermenter can be accurately grasped.

In the matrix data processing step, predictive analysis is performed on the matrix data using an algorithm simulating a human neural network, and bacterial flora data on two or more types of bacteria and data on the amount of biogas generated are included as variables. function can be obtained as prediction information.

Specific examples of predictive analysis using algorithms imitating human neural networks are not particularly limited, and include perceptrons, neural networks, convolutional neural networks, and deep learning. From the viewpoint of obtaining a more accurate function, predictive analysis using an algorithm that mimics a human neural network is preferably a perceptron, more preferably a neural network, more preferably a convolutional neural network, and more preferably deep learning. However, in the matrix data processing step, a highly accurate function can be obtained even in predictive analysis using an algorithm imitating a human neural network with low accuracy.

Predictive analytics using algorithms that mimic these human neural networks are well known. For example, the matrix data processing step can be performed according to the method described in "Toshiya Kaneshiro, Yumemiru Deep Learning Neural Network [Python Implementation] Introduction to Shuwa System". Said "Toshiya Kaneshiro, Yumemiru Deep Learning Neural Network [Python Implementation] Introduction to Shuwa System" is incorporated herein by reference. Predictive analysis using algorithms that mimic human neural networks can also be performed using commercially available software.

In the matrix data processing step, logistic regression analysis is performed on the matrix data, and a function containing bacterial flora data on two or more types of bacteria and data on the amount of biogas generated as variables can be obtained as prediction information.

Specific examples of logistic regression analysis are not particularly limited, and include coordinate descent method, gradient descent method, Newton method, and quasi-Newton method. From the viewpoint of obtaining a function with higher accuracy, the logistic regression analysis is preferably the coordinate descent method, more preferably the gradient descent method, more preferably the Newton method, and more preferably the quasi-Newton method. However, in the matrix data processing step, even logistic regression analysis with low accuracy can obtain a highly accurate function.

These logistic regression analyzes are well known. For example, the matrix data processing step can be performed according to the method described in "Toshiro Tango, Kazue Yamaoka, Haruyoshi Takagi, Logistic Regression Analysis - Actual Statistical Analysis Using SAS, Asakura Shoten". "Tshiro Tango, Kazue Yamaoka, Haruyoshi Takagi, Logistic Regression Analysis - Actual Statistical Analysis Using SAS, Asakura Shoten" is incorporated herein by reference. Logistic regression analysis can also be performed using commercially available software.

Although the function obtained by logistic regression analysis is not particularly limited, the following functions can be mentioned, for example. Note that the following function (the objective variable corresponds to the matrix x and the explanatory variable y corresponds to F(x)) is an example, and the present invention is not limited to the function;
F(x)=1/{1+e- ^(β0+β1x) } (function);
In the function, F(x) indicates data on the amount of biogas generated, and (β0+β1x) indicates microbial flora data of m types of bacteria. The explanatory variable y, corresponding to F(x), may be treated as 0-1 for logistic regression analysis and may be divided between 0 and 1 by some arbitrary threshold.

In the matrix data processing step, support vector machine analysis is performed on the matrix data, and a function containing bacterial flora data on two or more types of bacteria and data on the amount of biogas generated as variables can be obtained as prediction information. .

A specific example of support vector machine analysis is not particularly limited, and support vector regression can be mentioned, for example.

Support vector machine analysis is well known. For example, the matrix data processing step can be performed according to the method described in "Introduction to Support Vector Machines" by Nello Cristianini and John Shawe-Taylor, Kyoritsu Shuppan. The article "Nello Cristianini and John Shawe-Taylor, Introduction to Support Vector Machines, Kyoritsu Shuppan" is incorporated herein by reference. Support vector machine analysis can also be performed using commercially available software. Also, the matrix data processing step may be performed using programming using the R language or Python language, or Matlab using the C language, but the present invention is not limited to these.

<Accuracy confirmation process>
The accuracy confirmation step is a step that is performed after the matrix data processing step, and (i) substitutes the bacterial flora data A of the measurement data not included in the matrix data into a function to obtain a calculated value of the amount of biogas generated. and (ii) a comparison step of comparing the calculated value with the measured value of the amount of biogas that was being generated from the fermenter when the bacterial flora data A was obtained. .

The measurement data not included in the matrix data includes the bacterial lawn data A and the measured value of the amount of biogas generated from the fermenter when the bacterial lawn data A was obtained. These bacterial flora data A and the measured value of the amount of biogas are actually measured values obtained from an actually operated fermenter.

In the calculation step, the bacterial flora data A is substituted into the function to obtain a calculated value of the amount of biogas generated. Then, in the comparison step, by comparing the calculated value with the measured value of the amount of biogas generated from the fermenter when the bacterial flora data A was acquired, the accuracy of the function can be known.

Consider the relational expression “γ=|β−α|/β” where α is the calculated value and β is the measured value. For example, when "γ ≤ 0.5", more preferably when "γ ≤ 0.4", more preferably when "γ ≤ 0.3", more preferably when "γ ≤ 0.2" , More preferably, when "γ≦0.1", more preferably when "γ≦0.01", most preferably when "γ≦0.001", it is determined that the function has high accuracy. can be done.

<Other processes>
In the accuracy confirmation step described above, if the accuracy of the function is the desired accuracy, the function can be used, for example, in the processing method and processing system described later.

On the other hand, in the above-described accuracy confirmation step, if the accuracy of the function is not the desired accuracy, the function may be recreated, or the bacteria lawn data A and the bacteria lawn data that were not used to create the matrix data Matrix data may be newly created using the data of the amount of biogas generated when A was obtained. In this case, after the above-described accuracy confirmation step, (A) the work cycle including the matrix data creation step and the matrix data processing step; (B) the matrix data creation step, the matrix data processing step, and (C) a work cycle including a measurement data acquisition step, a matrix data creation step, and a matrix data processing step; and/or; (D) a measurement data acquisition step, a matrix data creation step, and a matrix A work cycle including a data processing step and an accuracy confirmation step may be performed. Note that each work cycle can be performed as many times as desired until the desired function is obtained. According to the above configuration, it is possible to obtain a function with higher accuracy.

[2. Processing method〕
The above [1. Prediction information creation method], it is possible to accurately predict the amount of biogas generated from actually acquired bacterial flora data in a plant in operation. The amount of biogas generated reflects the conditions in the fermenter. Therefore, if the amount of biogas generated can be accurately predicted, the state inside the fermenter can be accurately grasped. If the state in the fermenter can be accurately grasped, the state in the fermenter can be changed to a desired state (for example, a state in which the target substance in the fermenter can be efficiently processed by bacteria) as necessary. can do.

The processing method of the present embodiment is a processing method in which a target substance in a fermenter is treated with bacteria while generating biogas. Examples of the target object include biogas raw materials, waste water, bioenergy raw materials, and food raw materials. Examples of the biogas include methane gas, carbon dioxide, ammonia, hydrogen sulfide, hydrogen, water vapor, and acetic acid.

The processing method of the present embodiment, as shown in FIG. 2A, has at least a bacterial flora data acquisition step (S10) and a calculation step (S20). As shown in (B) to (D) of FIG. 2, the processing method of the present embodiment includes, in addition to the above steps, a state change step (S30) and/or a prediction information creation step (S40 ). Each step will be described below.

<Bacterial flora data acquisition process>
The bacterial flora data acquisition step is a step of acquiring bacterial flora data on two or more types of bacteria in the fermenter. The bacterial flora data can be obtained, for example, from a culture sample obtained from an operating plant or the like. The bacterial flora data is the above [1. Prediction information creation method]. Therefore, the flora data can be obtained based on the above function. The bacterial flora data acquisition step is performed by the bacterial flora data acquisition unit 2, which will be described later.

Bacterial flora data can be acquired by a known method. For example, the above [1. Prediction information creation method] can be used to obtain bacterial flora data. Therefore, the description of the method for acquiring bacterial flora data is omitted here.

Specific examples of the types of bacteria and microorganisms for which bacterial flora data is to be obtained are described in [1. Generating method of prediction information]. Therefore, descriptions of specific examples of bacteria and microorganisms are omitted here.

<Calculation process>
In the calculation step, bacterial flora data on two or more types of bacteria in the fermenter acquired in the bacterial flora data acquisition step is prepared in advance, bacterial flora data on two or more types of bacteria, and biogas. This is a step of calculating a predicted generation amount of biogas by substituting data of the generation amount into a function (for example, multiple regression equation) containing the data of the generation amount as a variable. The calculation process is performed by the calculation unit 3, which will be described later.

The functions used in the calculation process are those described in [1. Prediction information creation method]. Since the function is highly accurate, it is possible to accurately predict the amount of biogas generated in the calculation process. The amount of biogas generated reflects the conditions in the fermenter. Therefore, the calculation process provides an accurate understanding of the conditions within the fermenter. For example, if the predicted amount of biogas generated is small, it can be determined that the activity of bacteria in the fermenter is inactive (in other words, the fermentation phenomenon is inactive). If the generated amount is large, it can be determined that the activity of bacteria in the fermenter is active (in other words, the fermentation phenomenon is active).

<State change process>
The state change step is a step that can be performed after the calculation step and is a step that changes the conditions in the fermenter. By the state changing step, the state in the fermenter can be brought closer to a desired state (for example, a state in which the target substance in the fermenter can be efficiently processed by bacteria). The state changing process is performed by the state changing unit 5, which will be described later.

The changing conditions step can change the conditions in the fermenter in a variety of ways. For example, (i) adjustment of the amount of raw material (in other words, the target product to be treated by the fungus) to be put into the fermenter, (ii) composition of the raw material (VS concentration, TS concentration, or , C/N ratio); (vi) adjusting the amount of water introduced into the fermenter; (vii) adjusting the amount and/or type of nutrient added to the fermenter; (viii) Adjusting the amount and/or type of pH adjusting agent introduced into the fermenter, (ix) adjusting the agitation conditions in the fermenter, and/or (x) adjusting the residence time of the culture in the fermenter, etc. can change the conditions in the fermenter.

<Prediction information creation process>
The processing method of the present embodiment may have a prediction information creation step before the bacterial flora data acquisition step or between the bacterial flora data acquisition step and the calculation step. The prediction information creation step is a step of creating a function (for example, multiple regression equation) used in the calculation step. The prediction information creation step is performed by the prediction information creation unit 7, which will be described later. In addition, the prediction information creation unit 7, before actually operating the plant, based on the information obtained in advance from the fermenter 1 and the bacterial flora data acquisition unit 2, the function used in the calculation process (for example, multiple regression equation ) can be created.

The above [1. Prediction information creation method], it is possible to accurately predict the amount of biogas generated from actually acquired bacterial flora data in a plant in operation. The amount of biogas generated reflects the conditions in the fermenter. Therefore, if the amount of biogas generated can be accurately predicted, the state inside the fermenter can be accurately grasped. If the state in the fermenter can be accurately grasped, the state in the fermenter can be changed to a desired state (for example, a state in which the target substance in the fermenter can be efficiently processed by bacteria) as necessary. can do.

The prediction information creation step is the same as in [1. Prediction information creation method], and therefore the description of the prediction information creation process is omitted here. Note that the function (for example, multiple regression equation) used in the calculation step may be created separately from the processing method of the present embodiment.

[3. processing system]
The above [1. Prediction information creation method], it is possible to accurately predict the amount of biogas generated from actually acquired bacterial flora data in a plant in operation. The amount of biogas generated reflects the conditions in the fermenter. Therefore, if the amount of biogas generated can be accurately predicted, the state inside the fermenter can be accurately grasped. If the state in the fermenter can be accurately grasped, the state in the fermenter can be changed to a desired state (for example, a state in which the target substance in the fermenter can be efficiently processed by bacteria) as necessary. can do.

The processing system of the present embodiment is a processing system that processes a target object with bacteria while generating biogas. More specifically, in the processing system of the present embodiment, the above [2. processing method] is performed. Examples of the target object include biogas raw materials, waste water, bioenergy raw materials, and food raw materials. Examples of the biogas include methane gas, carbon dioxide, ammonia, hydrogen sulfide, hydrogen, water vapor, and acetic acid.

The processing system 10 of the present embodiment includes at least a fermenter 1, a bacterial flora data acquisition unit 2, and a calculation unit 3, as shown in FIG. Further, the processing system 10 of the present embodiment may further include a state changing section 5 and/or a prediction information creating section 7, as shown in FIG. Each configuration will be described below.

<Fermentation tank 1>
The fermenter 1 is for mixing the target substance and bacteria, and more specifically, it is for containing and mixing the target substance and bacteria to cause a fermentation phenomenon.

The fermenter 1 is not particularly limited as long as it can contain the target substance and bacteria. The fermentor 1 can be, for example, a vessel with a volume of 1 L or more, 10 L or more, 10 ² L or more, 10 ⁵ L or more, 10 ⁷ L or more, or 10 ¹⁰ L or more. The upper limit of the volume may be 10 ²⁰ L or 10 ¹⁵ L, for example.

The fermenter 1 may be provided with an opening through which a culture sample in the fermenter 1 can be obtained. A specific configuration of the opening is not limited. The opening has a cylindrical shape that extends to the area inside the fermenter 1 and to the lower area of the fermenter 1 (in other words, the area below the liquid surface of the culture existing in the fermenter 1). structures are connected. With this configuration, the culture is sucked or collected outside the fermenter 1 via the cylindrical structure, and as a result, the biogas in the fermenter can be obtained without contamination with outside air. can. The culture sample is sent to the bacterial flora data acquisition unit 2, which will be described later.

<Microbial lawn data acquisition unit 2>
The bacterial flora data acquisition unit 2 is for acquiring bacterial flora data regarding two or more types of bacteria in the fermenter 1 . The bacterial flora data can be obtained using culture samples obtained from the fermenter 1 . The bacterial flora data acquisition unit 2 performs the bacterial flora data acquisition step described above.

Bacterial flora data can be acquired by a known method. For example, the above [1. Prediction information creation method] can be used to obtain bacterial flora data. Therefore, the description of the method for acquiring bacterial flora data is omitted here.

Specific examples of the types of bacteria and microorganisms for which bacterial flora data is to be obtained are described in [1. Generating method of prediction information]. Therefore, descriptions of specific examples of bacteria and microorganisms are omitted here.

The bacterial flora data acquisition unit 2 is composed of a nucleic acid extraction device, a PCR device for probe amplification, and/or a device for acquiring bacterial flora data (eg, NGS, DGGE, RT-PCR, and/or microarray). can be A commercially available device can be used as each of the nucleic acid extraction device, the PCR device for probe amplification, and the device for acquiring bacterial flora data.

<Calculation unit 3>
The bacterial lawn data acquired by the bacterial lawn data acquisition unit 2 is sent to the calculation unit 3 . The calculation unit 3 substitutes the bacterial flora data into a previously created function (for example, a multiple regression equation) containing bacterial flora data on two or more types of bacteria and data on the amount of biogas generated as variables. By doing so, the predicted generation amount of biogas is calculated. The calculation unit 3 performs the calculation process described above.

The functions used in the calculator 3 (in other words, the functions stored in the calculator 3) are described in [1. Prediction information creation method]. Since the function is highly accurate, the calculation unit 3 can accurately predict the amount of biogas generated. The amount of biogas generated reflects the state inside the fermenter 1 . Therefore, the calculation unit 3 can accurately grasp the state inside the fermenter 1 . For example, if the predicted amount of biogas generated is small, it can be determined that the activity of bacteria in the fermenter 1 is inactive (in other words, the fermentation phenomenon is inactive), and the predicted biogas is generated, it can be determined that the activity of bacteria in the fermenter 1 is active (in other words, the fermentation phenomenon is active).

A specific configuration of the calculation unit 3 will be described later.

<State change unit 5>
The determination result of the calculation unit 3 (for example, the determination result that the activity of bacteria in the fermenter 1 is inactive or active) is sent to the state change unit 5 . The state changer 5 can change the state inside the fermenter 1 based on the determination result of the calculator 3 . The state change unit 5 performs the state change process described above.

The state changing unit 5 can change the state inside the fermenter 1 by various methods. For example, (i) adjustment of the amount of raw material (in other words, the target product to be treated by the fungus) to be put into the fermenter, (ii) composition of the raw material (VS concentration, TS concentration, or , C/N ratio); (vi) adjusting the amount of water introduced into the fermenter; (vii) adjusting the amount and/or type of nutrient added to the fermenter; (viii) Adjusting the amount and/or type of pH adjusting agent introduced into the fermenter, (ix) adjusting the agitation conditions in the fermenter, and/or (x) adjusting the residence time of the culture in the fermenter, etc. , the state in the fermenter 1 can be changed.

When changing the state in the fermenter 1 by the method (i) described above, the state changing unit 5 may, for example, include a channel for introducing raw materials into the fermenter 1 and/or a raw material in the channel. A pump (or valve) may be provided to supply the The flow path and/or pump (or valve) can be used to control the amount of raw material introduced into the fermenter 1 .

When the state inside the fermenter 1 is changed by the method (ii) described above, the state changing unit 5 may include, for example, a mixer for preparing raw materials to be introduced into the fermenter 1 . After adjusting the ratio of the plurality of raw materials charged into the mixer, the composition of the raw materials charged into the fermenter 1 can be adjusted by charging the plurality of raw materials into the fermenter 1 . In addition, the state changing unit 5 includes a plurality of flow paths for introducing each of the plurality of raw materials into the fermenter 1, and/or a pump (or , valves). By adjusting the amount of each raw material introduced into the fermenter 1 from each channel, the composition of the raw materials introduced into the fermenter 1 can be adjusted.

When the state inside the fermenter 1 is changed by the method (iii) described above, the state changing unit 5 may be provided with, for example, a heating device and/or a cooling device. The temperature in the fermenter 1 can be adjusted by heating and/or cooling the fermenter 1 and/or the raw materials introduced into the fermenter 1 by the heating device and/or the cooling device.

When changing the state in the fermenter 1 by the method (iv) described above, the state changing unit 5 is, for example, a channel for circulating the culture in the fermenter 1 and / or in the channel A pump (or valve) may be provided to supply the culture. The amount of culture circulating in the fermenter 1 can be regulated by the channel and/or pump (or valve).

When changing the state in the fermenter 1 by the above-described method (v), the state changing unit 5 may include, for example, a channel for taking out the culture from the fermenter 1 and/or a culture in the channel. A pump (or valve) may be provided to supply the The amount of culture removed from the fermentor 1 can be regulated by said channels and/or pumps (or valves).

When changing the state in the fermenter 1 by the above-described method (vi), the state changing unit 5 may, for example, include a flow path for introducing water into the fermenter 1 and/or water in the flow path. A pump (or valve) may be provided to supply the The amount of water introduced into the fermenter 1 can be regulated by the channel and/or pump (or valve).

When changing the state in the fermenter 1 by the method (vii) described above, the state changing unit 5 may include, for example, containers for individually accommodating various types of nutrient salts. By containing each of the various types of nutrients separately, the amount and/or type of nutrients fed into the fermentor 1 can be varied.

When changing the state in the fermenter 1 by the method (viii) described above, the state changing unit 5 may include, for example, containers for individually accommodating various types of pH adjusters. . By containing each of the various types of pH adjusters individually, the amount and/or type of pH adjuster fed into the fermentor 1 can be varied.

When the state inside the fermenter 1 is changed by the method (ix) described above, the state changing unit 5 may include, for example, a stirring device (for example, a stirring blade such as a propeller). The stirring conditions in the fermenter 1 can be adjusted by changing the operating state of the stirring device.

When changing the state in the fermenter 1 by the above-described method (x), the state changing unit 5 may include, for example, a channel for taking out the culture from the fermenter 1 and/or a culture in the channel. A pump (or valve) may be provided to supply the The residence time of the culture in the fermentor 1 can be adjusted by the channel and/or pump (or valve).

<Prediction information creation unit 7>
The processing system of this embodiment may include a prediction information creation unit 7 . The prediction information creating unit 7 creates a function (for example, multiple regression equation) to be used in the calculating unit 3 and sends the function to the calculating unit 3 . Of course, in the processing system of the present embodiment, the calculation unit 3 may store a function created with a configuration different from that of the prediction information creation unit 7 . The prediction information creation unit 7 performs the prediction information creation process described above.

The prediction information creation unit 7 performs the above [1. prediction information creation method]. For example, the prediction information creation unit 7 performs the above [1. prediction information creation method].

For example, biogas and a culture sample are obtained in the fermenter 1 in advance before the treatment system of the present embodiment is actually operated. The biogas is sent to the prediction information creation unit 7 and the culture sample is sent to the bacterial flora data acquisition unit 2 .

The prediction information creating unit 7 can acquire data on the amount of generated biogas using the sent biogas. In this case, the prediction information creation unit 7 may be equipped with a commercially available device for measuring the amount of specific gas generated.

On the other hand, the bacterial flora data acquisition unit 2 can acquire bacterial flora data using the sent culture sample. The bacterial flora data is sent to the prediction information creating section 7 .

The prediction information creation unit 7 has the same function as the bacterial flora data acquisition unit 2 , and both the biogas and the culture sample obtained in the fermenter 1 may be sent to the prediction information creation unit 7 . In this case, the prediction information creation unit 7 includes not only a commercially available device for measuring the amount of generated specific gas, but also a nucleic acid extraction device, a PCR device for probe amplification, and/or a device for acquiring bacterial flora data. (eg, NGS, DGGE, RT-PCR, and/or microarrays), and the like. A commercially available device can be used as each of the nucleic acid extraction device, the PCR device for probe amplification, and the device for acquiring bacterial flora data.

For example, before the processing system of the present embodiment is actually operated, a plurality of measurement data including biogas generation amount data and bacterial flora data are acquired in advance in the prediction information creation unit 7, and the measurement data Using at least two or more of the above, matrix data can be created with bacterial flora data as an explanatory variable and biogas generation amount data as an objective variable. Next, in the prediction information creation unit 7, analysis processing is performed on the matrix data, and a function (for example, multiple regression equation ) is obtained as prediction information.

Creation of matrix data, analysis processing, and the like may be performed by a specific configuration provided in the prediction information creation unit 7 . A specific example of the specific configuration will be described later.

<Configuration of Calculation Unit 3, and Example of Configuration Provided in Prediction Information Creation Unit 7 for Creating Matrix Data and Performing Analysis Processing (For example, Multiple Regression Analysis)>
The calculation unit 3 and the prediction information creation unit 7 (or part of the configuration of the calculation unit 3 and the prediction information creation unit 7) are realized by a logic circuit (hardware) formed in an integrated circuit (IC chip) or the like. Alternatively, it may be implemented by software.

In the latter case, the calculation unit 3 and the prediction information creation unit 7 (or part of the configuration of the calculation unit 3 and the prediction information creation unit 7) are provided with a computer that executes program instructions, which are software for realizing each function. ing. This computer includes, for example, at least one processor (control device) and at least one computer-readable recording medium storing the program. In the computer, the processor reads the program from the recording medium and executes it, thereby achieving the object of the present invention. As the processor, for example, a CPU (Central Processing Unit) can be used. As the recording medium, a "non-temporary tangible medium" such as a ROM (Read Only Memory), a tape, a disk, a card, a semiconductor memory, a programmable logic circuit, or the like can be used. In addition, a RAM (Random Access Memory) for developing the above program may be further provided. Also, the program may be supplied to the computer via any transmission medium (communication network, broadcast wave, etc.) capable of transmitting the program. Note that one aspect of the present invention can also be implemented in the form of a data signal embedded in a carrier wave in which the program is embodied by electronic transmission.

<1. Acquisition of bacterial flora data>
Methane fermentation was performed using a mixture of waste paper and garbage as raw materials. The operating conditions of the fermenter are as follows: (i) Fermenter A processed at an efficiency in which the amount of raw material converted to gas (gas conversion load) is less than 2.9 kg -VS/(m ³ d); (ii) a fermenter B treated at an efficiency with a gas equivalent load of 2.9 kg or more and less than 4.3 kg -VS/(m ³ ·d); (iii) a gas equivalent load of 4.3 kg or more5. Fermenter C treated at an efficiency of less than 7 kg - VS / (m ³ d), (iv) efficiency with a gas equivalent load of 5.7 kg or more and less than 7.1 kg - VS / (m ³ d) and (v) a fermenter E treated at an efficiency with a gas-equivalent load of 7.1 kg or more -VS/(m ³ ·d).

In addition to measuring the amount of methane gas generated from each fermenter, a culture sample (a mixture of raw materials and bacteria) was obtained from the fermenter at the time of measurement, and the bacteria contained in the culture sample (approximately 1276 types). bacteria) was acquired. As described above, measurement data including data on the amount of methane gas generated and bacterial flora data were obtained for a plurality of culture samples.

The amount of methane gas generated was measured using a well-known measuring device (specifically, a wet gas meter device manufactured by SINAGAWA Co., Ltd.). The specific measurement method followed the protocol attached to the measurement device.

Bacterial flora data was measured using a well-known measurement device (specifically, next-generation sequencer Miseq manufactured by Illumina Inc.). The specific measurement method followed the protocol attached to the measurement device.

<2. PCA Analysis and PLS Analysis>
PCA (Principal Components Analysis) analysis and PLS analysis using multivariate analysis software "pirouette 4.50" from Infometrix, with bacterial flora data as an explanatory variable and data on the amount of methane gas generated as an objective variable. did The specific methods of PCA analysis and PLS analysis followed the protocol attached to the multivariate analysis software "pirouette 4.50".

FIG. 5 shows the results of PCA analysis. In addition, in FIG. 5, the names of culture samples used for PCA analysis are described. In addition, in FIGS. 5 to 8, the names of culture samples derived from the same fermenter are written in letters of the same color tone (same density in black and white drawings). For example, a sample such as "2017_0714" is a sample derived from fermenter A, a sample such as "2017_0616" is a sample derived from fermenter B, a sample such as "2017_0929" is a sample derived from fermenter C, Samples such as "2017_0922" are samples derived from fermenter D, and samples such as "2017_0908" are samples derived from fermenter E.

From FIG. 5, the measurement data derived from the culture samples of each of the fermenters A to E are separated based on the first principal component (Factor 1) that expresses a lot of information, and the data of each of the fermenters A to E It was found that the measured data from the cultured samples were not effectively separated based on the second principal component compared to the first principal component. It could be said that much of the information in the bacterial flora data is related to the amount of gas generated that characterizes the fermenters A to E. Compared to the first principal component, the second principal component, which has less information, could express more differences between samples. In other words, although there are exceptions such as the sample names "2017_1006" and "2017_0908", it is possible to show a tendency for separation to occur based on the gas-equivalent load, which is the measurement data derived from the culture samples of each of the fermenters A to E. It was revealed.

In the PCA analysis, the measurement data derived from each culture sample of fermenters A to E showed a tendency to be separated based on the gas-equivalent load, so PLS analysis was performed using the measurement data of all culture samples. In PLS analysis, the accuracy of the prediction model (function) changes depending on the number of factors used. The greater the number of factors used, the higher the linearity of the prediction model (function). Decreased prediction accuracy. In this example, it was found that a well-balanced prediction model (function) can be obtained if the number of factors is about 5.

FIG. 6 shows the results of the PLS analysis. In FIG. 6, the horizontal axis indicates the gas conversion load calculated from the measured value (actual value) of the amount of methane gas generated, and the vertical axis indicates the predicted gas conversion calculated from the calculated value (predicted value) of the amount of methane gas generated. It shows the amount of load. Although there are some outliers such as the culture sample names "2017_0714" and "2017_0519", we were able to construct a generally good prediction model (function).

<3. Cross validation>
To check the accuracy of the prediction model (function), the measured data is divided into a training set and a test set, and the explanatory variables of the test set are substituted into the prediction model (function) constructed using the training set to obtain the calculated value of the objective variable. There is a method called cross-validation in which the calculated value is obtained and the measured value of the objective variable of the test set is compared.

Therefore, the outliers were removed from the measurement data, the measurement data after the outliers were removed were arbitrarily divided into a training set and a test set, and cross-validation was performed using the training set and the test set.

FIG. 7 shows the results of cross-validation. In FIG. 7, the culture sample names marked with a black mark correspond to the test set, and the culture sample names marked with a non-black mark correspond to the training set. . As is clear from FIG. 7, the prediction model (function) constructed using the training set was able to accurately predict the objective variable of the test set.

<4. Examination on Bacteria>
When bacteria contained in culture samples are compared, bacteria contained in one culture sample may not be contained in another culture sample. Therefore, among the 1276 species for which bacterial flora measurement data was detected, species that were commonly present in culture samples from almost all fermenters were selected. As a result, microorganisms belonging to the phylum Euryarchaeota, microorganisms belonging to the phylum Bacteroidetes, microorganisms belonging to the phylum Firmicutes, microorganisms belonging to the phylum OP9, and microorganisms belonging to the phylum Synergistetes , and successfully identified microorganisms belonging to the phylum Thermotogae. More specifically, these microorganisms were the following 21 types of microorganisms.

(1) phylum Euryarchaeota microorganisms belonging to the family Methanobacteriaceae; (2) phylum Euryarchaeota microorganisms belonging to the genus Methanothermobacter; (3) phylum Bacteroidetes Microorganisms belonging to the order Bacteroidales; (4) phylum Bacteroidetes microorganisms belonging to the family Porphyromonadaceae; (5) phylum Firmicutes microorganisms belonging to the genus Leuconostoc; (6) (7) Firmicutes Microorganisms belonging to the family Clostridiaceae; (8) Firmicutes Microorganisms belonging to the genus Clostridium; (9) (10) Microorganisms belonging to the genus Pelotomaculum; (11) Microorganisms belonging to the genus Ruminococcaceae; (12) Microorganisms belonging to the phylum Firmicutes family Ruminococcaceae; (13) Phylum Firmicutes microorganisms belonging to the genus Syntrophomonas; (14) Phylum Firmicutes (Sporomusa) Sporomusa (15) phylum Firmicutes microorganisms belonging to the phylum SHA-98; (16) phylum Firmicutes microorganisms belonging to the order Thermoanaerobacterales; (17) phylum Firmicutes Thermase belongs to the genus Thermacetogenium Microorganisms; (18) OP9 phylum Microorganisms belonging to the TIBD11 family; (19) Synergistetes phylum Microorganisms belonging to the Anaerobaculaceae family; (20) Synergistetes phylum Microorganisms belonging to the genus Anaerobaculum; 21) Microorganisms belonging to the genus S1 of the phylum Thermotogae.

Outliers were removed from the measurement data of these microorganisms, the measurement data after the removal of the outliers were arbitrarily divided into a training set and a test set, and cross-validation was performed using the training set and the test set.

FIG. 8 shows the results of cross-validation. In FIG. 8, the culture sample names marked with a black mark correspond to the test set, and the culture sample names marked with a black mark correspond to the training set. . As is clear from FIG. 8, the prediction model (function) constructed using the training set was able to accurately predict the objective variable of the test set.

<5. Summary>
We were able to construct a prediction model (function) that can predict the amount of methane gas generated, which is an index of the reaction state, using bacterial flora data. Since the amount of methane gas generated correlates with the amount of raw material to be processed, a prediction model (function) capable of predicting the amount of methane gas generated is also a prediction model (function) capable of predicting the amount of raw material to be processed. In addition, the bacterial lawn for which the bacterial lawn data has been acquired can process the raw material after the time of acquiring the bacterial lawn data. Therefore, the prediction model (function) described above can also be said to be a prediction model (function) capable of predicting reactions after acquisition of bacterial flora data.

INDUSTRIAL APPLICABILITY The present invention can be used in fields utilizing fermentation phenomena caused by microorganisms. For example, the present invention can be used in the fields of bioenergy production, wastewater treatment, and food production.

1 fermenter 2 bacterial flora data acquiring unit 3 calculating unit 5 state changing unit 7 prediction information creating unit 10 processing system S1 measurement data acquiring step S2 matrix data creating step S3 matrix data processing step S4 accuracy checking step S10 bacterial flora data acquiring step S20 calculation step S30 state change step S40 prediction information creation step

Claims (6)

A method for creating prediction information for predicting the amount of biogas generation, comprising:
When the data on the amount of biogas generated derived from a fermenter with a gas-equivalent load of 2.9 kg or more and less than 7.1 kg -VS/(m ³ d) and the data on the amount of biogas generated are obtained. a measurement data acquisition step of acquiring a plurality of measurement data including bacterial flora data on two or more types of bacteria in the fermenter;
a matrix data creation step of creating matrix data using at least two or more of the measurement data, using the bacterial flora data as an explanatory variable, and using the biogas generation amount data as an objective variable;
a matrix data processing step of performing analysis processing on the matrix data to obtain, as the prediction information, a function containing bacterial flora data on the two or more types of bacteria and data on the amount of biogas generated as variables; has
The above bacterial flora data includes (1) Euryarchaeota microorganisms belonging to the family Methanobacteriaceae; (2) Euryarchaeota microorganisms belonging to the genus Methanothermobacter; (4) Bacteroidetes, a microorganism belonging to the Porphyromonadaceae family; (5) Firmicutes, a microorganism belonging to the genus Leuconostoc Microorganisms belonging to; (6) phylum Firmicutes microorganisms belonging to the order Clostridiales; (7) phylum Firmicutes microorganisms belonging to the family Clostridiaceae; (8) phylum Firmicutes genus Clostridium (9) Microorganisms belonging to the Firmicutes phylum Lachnospiraceae family; (10) Firmicutes phylum Microorganisms belonging to the genus Pelotomaculum; (11) Firmicutes phylum Ruminococcaceae (12) Microorganisms belonging to the phylum Firmicutes family Ruminococcaceae; (13) Microorganisms belonging to the phylum Syntrophomonas genus; (14) Firmicutes (15) phylum (Sporomus) microorganisms belonging to the genus Sporomusa; (15) phylum Firmicutes microorganisms belonging to the phylum SHA-98; (16) phylum Firmicutes microorganisms belonging to the order Thermoanaerobacterales; (Firmicutes) Microorganisms belonging to the genus Thermacetogenium; (18) OP9 phylum Microorganisms belonging to the TIBD11 family; (19) Synergistetes phylum Microorganisms belonging to the Anaerobaculaceae family; (20) Synergistetes phylum Microorganisms belonging to the genus Anaerobaculum; (Thermotogae) includes bacterial flora data of two or more types of microorganisms selected from the group consisting of microorganisms belonging to the genus S1 ,
In the matrix data processing step, multiple regression analysis is performed on the matrix data, and the multiple regression equation including the bacterial flora data on the two or more types of bacteria and the data on the amount of biogas generated as variables is predicted. A method of creating prediction information, which is a process of obtaining information . After the matrix data processing step, (i) a calculation step of substituting the bacterial flora data A of the measurement data not included in the matrix data into the function to obtain a calculated value of the amount of biogas generated; ) an accuracy confirmation step including a comparison step of comparing the calculated value with a measured value of the amount of biogas generated from the fermenter when the bacterial flora data A was acquired. The preparation method described in . A treatment method for treating a target object in a fermenter with bacteria while generating biogas,
a bacterial flora data acquisition step of acquiring bacterial flora data on two or more types of bacteria in the fermenter having a gas-equivalent load of 2.9 kg or more and less than 7.1 kg -VS/(m ³ d);
The predicted amount of biogas generated is obtained by substituting the bacterial flora data into a previously created function containing the bacterial flora data on the two or more types of bacteria and the data on the amount of biogas generated as variables. and a calculating step of calculating
The above bacterial flora data includes (1) Euryarchaeota microorganisms belonging to the family Methanobacteriaceae; (2) Euryarchaeota microorganisms belonging to the genus Methanothermobacter; (4) Bacteroidetes, a microorganism belonging to the Porphyromonadaceae family; (5) Firmicutes, a microorganism belonging to the genus Leuconostoc Microorganisms belonging to; (6) phylum Firmicutes microorganisms belonging to the order Clostridiales; (7) phylum Firmicutes microorganisms belonging to the family Clostridiaceae; (8) phylum Firmicutes genus Clostridium (9) Microorganisms belonging to the Firmicutes phylum Lachnospiraceae family; (10) Firmicutes phylum Microorganisms belonging to the genus Pelotomaculum; (11) Firmicutes phylum Ruminococcaceae (12) Microorganisms belonging to the phylum Firmicutes family Ruminococcaceae; (13) Microorganisms belonging to the phylum Syntrophomonas genus; (14) Firmicutes (15) phylum (Sporomus) microorganisms belonging to the genus Sporomusa; (15) phylum Firmicutes microorganisms belonging to the phylum SHA-98; (16) phylum Firmicutes microorganisms belonging to the order Thermoanaerobacterales; (Firmicutes) Microorganisms belonging to the genus Thermacetogenium; (18) OP9 phylum Microorganisms belonging to the TIBD11 family; (19) Synergistetes phylum Microorganisms belonging to the Anaerobaculaceae family; (20) Synergistetes phylum Microorganisms belonging to the genus Anaerobaculum; (Thermotogae) includes bacterial flora data of two or more types of microorganisms selected from the group consisting of microorganisms belonging to the genus S1 ,
The calculation step includes substituting the bacterial flora data into a multiple regression equation that includes the previously created bacterial flora data on the two or more types of bacteria and the data on the amount of biogas generated as variables. , a processing method, which is a step of calculating a predicted generation amount of biogas . 4. The processing method according to claim 3 , further comprising, after the calculating step, a state changing step of changing the state of the target object in the fermenter to a state that can be processed by the fungus. A treatment system that treats an object with bacteria while generating biogas,
a fermenter for mixing the object and the bacteria;
a bacterial flora data acquisition unit for acquiring bacterial flora data on two or more types of bacteria in the fermenter having a gas-equivalent load of 2.9 kg or more and less than 7.1 kg -VS/(m ³ d);
The predicted amount of biogas generated is obtained by substituting the bacterial flora data into a previously created function containing the bacterial flora data on the two or more types of bacteria and the data on the amount of biogas generated as variables. and a calculation unit for calculating
The above bacterial flora data includes (1) Euryarchaeota microorganisms belonging to the family Methanobacteriaceae; (2) Euryarchaeota microorganisms belonging to the genus Methanothermobacter; (4) Bacteroidetes, a microorganism belonging to the Porphyromonadaceae family; (5) Firmicutes, a microorganism belonging to the genus Leuconostoc Microorganisms belonging to; (6) phylum Firmicutes microorganisms belonging to the order Clostridiales; (7) phylum Firmicutes microorganisms belonging to the family Clostridiaceae; (8) phylum Firmicutes genus Clostridium (9) Microorganisms belonging to the Firmicutes phylum Lachnospiraceae family; (10) Firmicutes phylum Microorganisms belonging to the genus Pelotomaculum; (11) Firmicutes phylum Ruminococcaceae (12) Microorganisms belonging to the phylum Firmicutes family Ruminococcaceae; (13) Microorganisms belonging to the phylum Syntrophomonas genus; (14) Firmicutes (15) phylum (Sporomus) microorganisms belonging to the genus Sporomusa; (15) phylum Firmicutes microorganisms belonging to the phylum SHA-98; (16) phylum Firmicutes microorganisms belonging to the order Thermoanaerobacterales; (Firmicutes) Microorganisms belonging to the genus Thermacetogenium; (18) OP9 phylum Microorganisms belonging to the TIBD11 family; (19) Synergistetes phylum Microorganisms belonging to the Anaerobaculaceae family; (20) Synergistetes phylum Microorganisms belonging to the genus Anaerobaculum; (Thermotogae) includes bacterial flora data of two or more types of microorganisms selected from the group consisting of microorganisms belonging to the genus S1 ,
The calculation unit substitutes the bacterial flora data into a multiple regression equation that includes, as variables, the bacterial flora data related to the two or more types of bacteria and the data of the amount of biogas generated, which has been created in advance. , a treatment system for calculating the expected generation of biogas . 6. The processing system according to claim 5 , further comprising a state changing unit that changes the state of the object in the fermenter to a state that can be processed by the bacteria.

Images