Patent Application
20250225486
This disclosure relates to an evaluation value calculation device for biogas materials, wherein the device is to match a source of food waste such as restaurants and food factories (biogas materials) with a biogas plant that performs methane fermentation and methane gasification using food waste.
Recently the problem of Global warming has become more and more serious. Efforts have been made to reduce the emission of greenhouse gas such as carbon dioxide gas and methane gas. The Japanese Ministry of the Environment has set a goal of achieving carbon neutrality, where CO2 (carbon dioxide) emissions will be reduced to practically zero by 2050, in order to achieve a decarbonized society (for example, refer to Non-Patent Document 1).
There are various ways to reduce CO2, such as the introduction of renewable energy equipment like solar and wind power generation, the introduction of energy-saving products such as heat pumps, and forest management activities. Recycling biomass (recycling food waste) is also thought to be one of them.
The primary sources of food waste are households and businesses. Food waste generated by businesses is classified as business waste and treated as industrial waste.
A biomass generation system has been developed to convert waste oil, previously treated solely as industrial sludge, into original biomass fuels using a diesel generator. This system represents a significant advancement by enabling CO2 reduction, recycling, and water purification. Relevant technologies and applications are detailed in Patent Documents 1 to 3.
On the other hand, challenges persist in promoting the recycling of “residual sludge (1002)” and “food loss (1003)” among business-related food waste. Table 1100 in
Currently, major food and beverage chains such as McDonald's® and Kentucky Fried Chicken® dispose of these food wastes as industrial waste, leaving them untapped as resources for energy generation.
The initial process in converting these unused food wastes and sludge into valuable resources involves processing them into biogas (e.g., methane for power generation or green hydrogen) at a biogas plant utilizing methane fermentation. Notably, in the context of methane power generation, the amount of biogas produced per ton of raw material from business-related food waste (including residual sludge) is 5 to 10 times greater than that produced from livestock manure or sewage sludge, which have traditionally served as the primary raw materials for conventional methane gas production. Consequently, methane power generation through the treatment of food residues, either alone or in combination with livestock manure or other materials, is strongly recommended.
However, due to the unstable properties of food waste, the operational efficiency of biogas plants remains low, ranging from 38% to 49%. While there is a recognized need to recycle food waste as a resource, progress in methane utilization has been constrained by specific challenges, detailed below.
(A) First, pre-diagnosing the unstable properties of food waste is a critical step before processing it at a biogas plant. However, the cost of current services for estimating gas generation is prohibitively high, ranging from 200,000 to 500,000 yen per sample. Additionally, seasonal variations in the volume of discharged food waste and the difficulties associated with diagnosing small-lot food waste in advance present significant operational challenges.
Due to these factors, biogas plants face difficulties in accurately predicting gas production. As a result, they prioritize avoiding fermentation defects and maintaining stable operations over incorporating new raw materials. This dependency on a limited range of raw materials has contributed to the low utilization rates of biogas plants.
(B) Second, conducting advanced evaluations of gas generation potential and the operational behavior of biogas plants, particularly when existing primary raw materials such as livestock manure are combined with new commercial food waste (mixed biogas materials), could significantly enhance operation management. However, the accuracy of current preliminary evaluations for mixed biogas materials remains insufficient. This shortcoming directly limits the recycling of residual sludge and food loss, thereby obstructing their effective utilization as valuable resources.
(C) Third, insufficient or imbalanced levels of nutrients (such as sulfur, nitrogen, and oil) required for methane bacteria to ferment, or a deficiency in lipase, an enzyme that facilitates the decomposition of cellulose, can result in inadequate fermentation. Consequently, the amount of gas generated may fall below expectations.
Currently, biogas plants lack the capability to determine optimal methods for promoting gasification to enhance gas generation. They are also unable to accurately identify and apply the appropriate type and quantity of gasification accelerators. This limitation contributes to reduced gas generation and, ultimately, a low capacity utilization rate.
The present application addresses the aforementioned challenges and aims to provide an evaluation value calculation device for biogas materials that promotes the active utilization of biogas resources by presenting specific evaluation values. As a result, this application resolves (A) the lack of a prior diagnostic mechanism for biogas materials and (B) and (C) the absence of evaluation values necessary for the stable operation of power plants when biogas materials are introduced, which have been obstacles to the utilization of food waste and other materials as biogas feedstocks.
To address the aforementioned issues, the present application provides an evaluation value calculation device for biogas materials that calculates an evaluation value related to the resource recovery of biogas materials. The device includes an input unit that accepts inputs regarding a machine learning library related to biogas materials or property data concerning biogas materials, and a machine learning model that receives the machine learning library or property data from the input unit. It further comprises a learning process execution unit that trains the machine learning model based on the machine learning library and an evaluation value calculation execution unit that estimates an evaluation value related to the resource recovery of biogas materials by performing an estimation process based on the property data while utilizing the machine learning model. The property data includes at least one of the following attributes: type, collection location, property, and quantity of biomass materials.
In the evaluation value calculation device, it is optional that data regarding the operational status of a biogas plant is input via the input unit. The data regarding the operational status of the biogas plant is then further provided to the machine learning model. The evaluation value calculation execution unit estimates an evaluation value related to the resource recovery of biogas materials by performing an estimation process based on both the property data and the data regarding the operational status of the biogas plant, utilizing the machine learning model. The data regarding the operational status of the biogas plant includes at least one of the following: the type of biogas material used at the plant. the fermentation method, the presence and amount of additives such as coagulants, the VS (Volatile Solids) volume load per 1 m3 of the methane fermentation tank, the organic decomposition rate in the methane fermentation tank, the methane gas generation amount in the gas holder, and the measured methane yield (BMP).
In the evaluation value calculation device, it is optional that the evaluation value estimated by the evaluation value calculation execution unit includes at least one of the following parameters related to mixed biogas materials, where an individual biogas material or a plurality of individual materials are combined:
In the evaluation value calculation device, it is optional that the evaluation value calculation execution unit estimates at least one of the following parameters regarding mixed biogas materials:
In the evaluation value calculation device, it is optional that the evaluation value calculation execution unit determines that the compatibility of mixed fermentation for mixed biogas materials improves as the methane yield in parameter (g) increases, provided that all conditions in parameters (a) to (f) are satisfied.
In the evaluation value calculation device, it is optional that the evaluation value calculation execution unit determines that a combination is unsuitable for mixed fermentation of biogas materials if the methane yield is less than 100 mL/g-VS.
In the evaluation value calculation device, it is optional for the evaluation value calculation execution unit to additionally determine the following conditions:
In the evaluation value calculation device, it is optional that the evaluation value calculation execution unit further predicts and analyzes the appropriate types and quantities of gasification promoters, nutritional supplements, and auxiliary raw materials.
In the evaluation value calculation device, it is optional that the evaluation value calculation execution unit further estimates the following as the evaluation value: whether pretreatment of each biogas material is necessary; the organic substance composition of each methane material; fermentation temperature and duration; organic substance concentration (VS) of each methane material; solid substance amount (TS); organic substance ratio (VS/TS); ash concentration; chemical oxygen demand (COD); theoretical oxygen demand (ThOD); COD/VS value; organic substance constituent element concentrations of each methane material, including the content of carbon (C: % VS), nitrogen (N: % VS), oxygen (O: % VS), and hydrogen (H: % VS); and the hemicellulose, cellulose, and lignin concentrations (% VS) of each methane material.
In the evaluation value calculation device, it is optional that the machine learning library includes at least one of the following: whether pretreatment of individual biogas materials and mixed biogas materials is necessary; the organic substance composition of each methane material; fermentation temperature and duration; organic substance concentration (VS); solid substance amount (TS); organic substance ratio (VS/TS); ash concentration; chemical oxygen demand (COD); theoretical oxygen demand (ThOD); COD/VS value; the concentrations of organic substance constituent elements in each methane material, including carbon (C: % VS), nitrogen (N: % VS), oxygen (O: % VS), and hydrogen (H: % VS); the concentrations of nutrients such as carbohydrates, lipids, proteins, and ash in each methane material; the concentrations of hemicellulose, cellulose, and lignin (% VS) in each methane material; the C/N ratio (ratio of carbon concentration to nitrogen concentration in each methane material); the lipid/VS ratio (ratio of lipid concentration to VS concentration in each methane material); Na concentration; the concentrations of micronutrients, including iron (mg/kg-COD raw material), cobalt (mg/g-COD raw material), and nickel (mg/g-COD raw material) when materials are mixed; theoretical methane yield (ThBMP: ml/g-VS); actual methane yield (BMP: ml/g-VS); predicted values of biodegradation rate, VS decomposition rate, and energy yield.
In the evaluation value calculation device, it is optional to further include a communication unit that enables a connection to a communication network, wherein the evaluation value related to the resource recovery of biogas materials, estimated by the machine learning library and the evaluation value calculation execution unit, is transmitted via the communication unit.
To address the aforementioned challenges, a program for an evaluation value calculation device is provided, which calculates an evaluation value related to the resource recovery of biogas materials. The program includes an input step, which involves receiving inputs regarding a machine learning library related to biogas materials or property data concerning biogas materials and inputting the machine learning library or property data into a machine learning model. It also includes a learning process execution step, which trains the machine learning model using the machine learning library to determine setting values. Furthermore, the program includes an evaluation value calculation execution step, which estimates an evaluation value related to the resource recovery of biogas materials by performing an estimation process based on the property data while utilizing the machine learning model. The property data includes at least one of the following attributes: type, collection location, property, and quantity of biomass materials.
In this program, it is optional that data regarding the operational status of a biogas plant is input during the input step, and that the data regarding the operational status of the biogas plant is provided to the machine learning model. The evaluation value related to the resource recovery of biogas materials is then estimated by performing an estimation process based on the property data and the operational status data of the biogas plant, utilizing the machine learning model during the evaluation value calculation execution step. The data regarding the operational status of the biogas plant includes at least one of the following: the type of biogas material used at the plant, the fermentation method, the presence and amount of additives such as coagulants, the VS volume load per 1 m3 of the methane fermentation tank, the organic decomposition rate in the methane fermentation tank, the methane gas generation amount in a gas holder, and the measured methane yield (BMP).
To address the aforementioned challenges, an evaluation value calculation method is provided for calculating an evaluation value related to the resource recovery of biogas materials. The method includes an input step for receiving inputs regarding a machine learning library related to biogas materials or property data concerning biogas materials and for inputting the machine learning library or property data into a machine learning model. It also includes a learning process execution step for training the machine learning model based on the machine learning library to determine setting values, and an evaluation value calculation execution step for estimating an evaluation value related to the resource recovery of biogas materials by performing an estimation process based on the property data while utilizing the machine learning model. The property data includes at least one of the following: type, collection location, property, and quantity of biomass materials.
In the evaluation value calculation method, it is optional that data regarding the operational status of a biogas plant is input during the input step and provided to the machine learning model. An evaluation value related to the resource recovery of biogas materials is then estimated during the evaluation value calculation execution step by performing an estimation process based on both the property data and the operational status data of the biogas plant, utilizing the machine learning model. The data regarding the operational status of the biogas plant includes at least one of the following: the type of biogas material used at the plant, the fermentation method, the presence and amount of additives such as coagulants, the VS volume load per 1 m3 of a methane fermentation tank, the organic decomposition rate in the methane fermentation tank, the methane gas generation amount in a gas holder, and the measured methane yield (BMP).
This application pertains to an evaluation value calculation device that calculates evaluation values for biogas materials. The device includes an input unit that receives inputs related to a machine learning library concerning biogas materials or property data regarding biogas materials, and a machine learning model that processes the machine learning library or property data received from the input unit. It also comprises a learning process execution unit that trains the machine learning model based on the machine learning library to determine setting values, and an evaluation value calculation execution unit that estimates evaluation values related to the resource recovery of biogas materials by performing estimation processes based on the property data while utilizing the machine learning model. The property data includes at least one of the following attributes: type, collection location, property, and quantity of biomass materials. With this configuration, the evaluation value calculation device for biogas materials enables the optimal matching of food waste sources, such as residual sludge and food loss, with biogas plants that perform methane gasification using food waste. This improves the operational efficiency of biogas plants and contributes to the realization of a decarbonized society in the long term.
A methane resource matching system, incorporating an evaluation value calculation device for biogas materials according to an embodiment of the present application, is described with reference to the accompanying figures. This methane resource matching system forms part of a decarbonization resource recovery service that supplies optimal biogas materials, such as food waste, from sources including restaurant chains and other contributors, based on the operational status of biogas plant operators.
In this embodiment, biogas materials include food waste, residual sludge (including refined sludge), food loss, and livestock waste, but exclude wastewater oil. Sources of biogas materials include, but are not limited to, restaurant chains, food factories, commercial facilities, supermarkets, convenience stores, hotels, schools, hospitals, and employee cafeterias.
First, an overview of the overall process for a methane resource recovery service, which includes the methane resource matching system (hereinafter referred to as the “matching system”), is provided with reference to
As the first step, in the methane material collection process, residual sludge (sedimentation residue), food loss, food waste (including lipid-based waste), and refined sludge are collected from biomass discharged by sources such as restaurants and food factories. Here, biomass refers to “renewable organic resources derived from living organisms, excluding fossil resources.” Among biomass, waste biomass includes discarded paper, livestock manure, food waste, and sewage sludge.
The CO2 emitted during the combustion of biomass is equivalent to the CO2 absorbed from the atmosphere through photosynthesis during the growth of living organisms. Therefore, replacing fossil resource-based energy and products with biomass significantly contributes to reducing CO2 emissions, where CO2 is one of the greenhouse gases responsible for global warming.
In general, significant amounts of food loss, including unsold products, expired items, and leftover ingredients, are generated daily from food waste sources such as food service establishments, fast food restaurants, restaurants, hotels, and food processing plants. These materials are collected as methane resources.
Additionally, wastewater from food waste sources contains various water pollutants. If such wastewater is discharged without proper treatment, substances such as oil in the wastewater can adhere to and solidify within drain pipes, causing blockages. This not only makes water purification in combined treatment tanks and sewage treatment plants more difficult but also negatively impacts the environment.
Therefore, facilities for dumping wastewater, including solid substances such as floating substances, comprise processing units (grease traps), oil/water separation tanks, and raw water tanks in order to pool wastewater for the time being or to deposit solid substances, or to float and separate oil, or to remove these temporarily for physical removal.
In the methane material collecting process, a field agent, which is, for example, a working company, manages the cleaning of grease traps at restaurants, etc., and collects waste oil and residual sludge from the grease traps. Herein, the device disclosed in U.S. Pat. No. 4,401,007, the device disclosed in U.S. Pat. No. 4,420,750, and the device disclosed in utility U.S. Pat. No. 3,216,173, are applicable as a technology to collect biomass from a grease trap. The applicant owns these patents.
Other field agents are, for example, agents that manage oil/water separation tanks at food factories that produce box lunches, frozen foods, and meat products, and collect biomass from the oil/water separation tanks. Specifically, oil sludge obtained from the oil/water separation tank at a food factory is separated into oil, water, and sludge using an oil sludge separation apparatus, and the sludge (residual sludge) is collected. The field agent may also manage the oil sludge separation apparatus.
Next, in the matching process, food waste (biomass materials such as food loss and so on) collected from restaurants, food factories, etc. in the methane material collecting process, and a biogas plant that uses this food waste to produce methane gas, are optimally connected using IoT (Internet of Things) and AI (Artificial Intelligence) in order to improve equipment utilization rate and productivity at biogas plants.
In this matching process, for example, food waste that has been treated as waste, presents the types, mixing ratios, and quantities of food waste that can serve as an appropriate methane resource according to conditions such as requests and operational status (for example, methane resources used and power generation status) from biomass plants. Collection and supply services can be provided. This will provide food-related businesses with various values such as waste reduction, decarbonization, and cost reductions. In addition, by supplying the optimal methane materials to individual biogas plants, values such as stabilizing gas volume and fermentation, and reducing operating costs will be created. The amount of electricity generated will increase in proportion to the increase in gas production, which can contribute to a decarbonization.
Next, in the methane gas production/consumption process, the biogas plant operator produces biogas (methane gas) from the biomass (livestock manure, food waste, oil residue (including purified sludge), food loss, and so on) brought into the biogas plant after the above matching process. Said biogas is generated by fermenting livestock manure, food loss, and residual sludge in a fermentation tank, and is generally stored in a gas holder and then used for power generation. In the future, it is expected that this methane gas will be reformed and used as green hydrogen based on non-fossil fuels.
Herein, businesses, that have introduced a biofuel (methane gas) produced from food waste, are entitled to receive carbon credits. In a case when a business has a CO2 reduction target, the amount more than the target will become CO2 reduction credits. In a case when a small business entity does not have any CO2 reduction target, all amounts will become saved as carbon credits. In Japan, said carbon credit can be sold via a J-credit system and can be sold overseas using the carbon credit trading system established by each country.
Next, the overall system of the matching system S, according to this embodiment, will be described with reference to
The terminal device (20) is a portable device, such as a personal computer or smartphone, used at a biogas plant to process food waste into biogas (also referred to as methane gasification or methane resource recovery). The terminal device (20) is equipped with a specialized application designed for field investigation and matching purposes, allowing users to input basic data related to methane gasification. The results of the input and analysis are displayed on the device's screen.
The terminal device (20) establishes a communication session with the management server (50) using a specified protocol and facilitates the transmission and reception of data, files, and other information related to the management of methane materials, methane gas production, and methane power generation at the biogas plant. Additionally, the terminal device (20) can send requests for specific methane materials to the management server (50) in formats such as HTTP requests.
The terminal device (30) is a device, such as a smartphone, used at the site of the food waste source or at the head office of the source to input information regarding the food waste to be discharged. The terminal device (30) is equipped with a specialized application designed for source investigation and matching purposes. This application allows users to input basic matching data for methane resource recovery and displays the results on the device's screen. The specialized application may also include a web-based visualization tool for managing the collection of food waste to be discharged.
The terminal device (30) establishes a communication session with the terminal device (40) at the support company and the management server (50) using a predetermined protocol. It facilitates the transmission and reception of photos of food waste, as well as data and files related to the management of food waste and wastewater oils and fats.
The terminal device (40) is utilized at a support company responsible for investigating and collecting food waste. It is a device, such as a smartphone or personal computer, used to input information regarding food waste at its source, the raw materials used, and the operating status of biogas plants. The terminal device (40) runs a specialized application and displays the results on its screen.
Using a predetermined protocol, the terminal device (40) facilitates the transmission and reception of data and files between the terminal devices (20 and 30) and the management server (50). Additionally, the terminal device (40) acquires information about newly discharged food waste from the terminal device (30) and information requested by the biogas plant from the terminal device (20), often in the form of HTTP requests or similar formats. Based on optimal methanation material matching, the terminal device (40) efficiently matches the food waste source with the biogas plant.
The management server (50) is located at the general headquarters responsible for overseeing the bio-gasification of food waste. It manages the bio-gasification processes at both the sources of food waste and the biogas plants. Specifically, the management server (50) obtains information about raw materials used and operating conditions at biogas plants from terminal devices (20 or 30) and about food waste discharged from terminal devices (30 or 40). Utilizing this information, the management server (50) efficiently facilitates methane resource recovery.
In this process, the management server (50) centrally manages the initial matching service, resource recovery service, and CO2 reduction management service. These services are primarily executed through the four programs (401 to 404) shown in
The IoT source investigation program (401) identifies the type, quantity, and properties of methane materials managed by sources of food waste based on input data and photographs sent via a smartphone application or similar means from terminal devices (30 or 40). The IoT source investigation program (401) also determines the type and quantity of methane materials required by the methane gas plant based on gas production levels, power generation status, and requests sent from the biogas plant's terminal devices (20 or 30).
In other words, the management server (50), using the IoT source investigation program (401), collects and integrates data from terminal devices (20, 30, and 40) regarding source investigations, gas production levels, and power generation status at biogas plants. By consolidating data that was previously dispersed, such as food waste data from sources including food factories and gas production or power generation data from biogas plants, the management server (50) enhances the efficiency and productivity of methane gas generation and methane power generation.
Specifically, mobile devices such as smartphones (terminal devices 30, 20, and 40) are utilized at food waste sources, biogas plants, and support companies, respectively. Data from these devices can be instantly uploaded to the cloud, enabling the use of big data for further analysis without requiring physical visits to the sources.
The management server (50), using the IoT source investigation program (401), monitors and evaluates the status of food waste sources and biogas plants. As a result, the time required for processing data related to source investigations, which previously took 30 minutes per site, can be reduced to just 1 minute per site with the implementation of this program.
The AI methane resource recovery program (402) is used by the management server (50) and related systems to predict methane yield for individual materials and diagnose inhibition effects when multiple materials are mixed with existing raw materials. These predictions are based on past research data, newly collected analytical data, actual operational data, and other sources.
In this context, methane yield refers to the estimated amount of methane produced per unit of decomposed volatile solids (VS), which represent the organic content of food waste as a raw material. Additionally, synergy prediction refers to forecasting the synergistic effects on methane yield when different types of methane materials are mixed.
With the use of the AI methane resource recovery program (402), the management server (50) constructs a highly accurate prediction model and applies precise methane yield estimation formulas tailored for individual materials. Additionally, estimation formulas incorporating new analytical methods, nutrient factors, inhibition caused by oils and fats, and synergies resulting from raw material mixtures are developed and implemented.
This centralized approach allows the management server (50) at headquarters to efficiently manage the evaluation of energy conversion equipment and raw materials used in methane fermentation.
Specifically, in determining individual materials, methane yield and other fundamental data (actual data) are stored in a learning database for each raw material, such as food waste (including lipid-based waste), residual sludge (including refined sludge), food loss, and livestock waste used for methane fermentation. Machine learning is then performed through mathematical calculations and corrections to actual measured values, resulting in the creation of big data.
The implementation of the AI methane resource recovery program (402) enables highly accurate estimation of the potential for methane resource recovery for both individual materials and mixed materials. As a result, the basic determination of methane resource recovery potential, which previously required 180 days per raw material, can now be completed in just 2 seconds per raw material using the AI methane resource recovery program (402). Details regarding the AI methane resource recovery program (402) will be provided later in this document.
The management server (50), through the implementation of the AI gasification promotion program (403), identifies the optimal gasification promotion method and forecasts the appropriate input amounts and expected effects of gasification promoters, including nitrogen, sulfur, oils and fats, and lipase. These processes aim to resolve estimated inhibiting factors and optimize the utilization of materials with low methane yield.
Specifically, the AI gasification promotion program (403) is utilized to identify the optimal gasification promotion method based on the type, quantity, and properties of food waste. The program calculates the raw material mixing ratio and determines the optimal dosage of accelerators, such as purified sludge, glycerin, and rumen bacteria. Test data for each mixing ratio, fundamental data for each gasification accelerator candidate, and additional data from administration tests are stored in the AI learning database.
Machine learning and deep learning are then performed, incorporating mathematical calculations and corrections to actual measured values, while simultaneously generating big data. As a result, the optimal gasification promotion method, raw material mixing ratio, and dosage for each gasification accelerator can be determined with high precision. Previously, these processes required 180 days, but with the implementation of AI, they can now be completed in just 10 seconds per raw material.
Gasification promoters primarily consist of auxiliary raw materials, nutritional supplements, and accelerators. Among these, auxiliary raw materials do not promote the decomposition of other raw materials. However, because auxiliary raw materials exhibit a high methane yield per unit weight, mixing them with other raw materials can increase the apparent methane yield per unit of input material.
It should be noted, however, that auxiliary raw materials are generally difficult to process through methane fermentation when used as a sole raw material. Examples of auxiliary raw materials include refinery sludge from grease trap oil and biodiesel fuel (BDF) residue glycerin. Methane gasification can be significantly enhanced when the mixing ratios of auxiliary raw materials and food waste are properly optimized.
Nutritional supplements, including nickel (Ni), cobalt (Co), and iron (Fe), are essential for the metabolism of methanogens but are typically insufficient in most raw materials. The addition of these supplements enhances nutrient availability and accelerates the fermentation process, particularly in food waste.
For example, nutritional supplements may include compounds such as NiCl2·6H2O, CoCl2·6H2O, and FeCl2·4H2O, which are already available in commercial nutrient solutions. However, the optimal amount of these supplements depends on specific conditions, and their effective application remains a challenge in current practices.
Some studies report that promoters are effective in decomposing cellulose. These promoters have the potential to enhance methane yields from cellulose-rich agricultural residues, such as resource crops like rice straw; rice husks, and dent corm. However, promoters may not significantly increase gas production from food-based waste.
Furthermore, conducting the enzyme reaction requires separate reaction tank equipment, with a reaction time ranging from several hours to one day. Examples of promoters include cellulase enzymes and rumen fluid.
As illustrated in Table 400 of
After machine learning, the AI gasification promotion program (403) determines the optimal gasification promotion method based on the type, quantity, and properties of the food waste. It also calculates the raw material mixing ratio and automatically determines both the type and optimal input amount of the gasification promoter.
The management server (50), through the implementation of the CO2 reduction IT calculation program (404), estimates the amount of methane gas generated and the corresponding CO2 reduction. Additionally, the program calculates the CO2 reduction volume and the amount that can be converted into carbon credits based on the actual utilization of biogas at biogas plants.
The management server (50) centrally manages data (via a cloud platform) related to the collection volumes of “residual sludge (including refined sludge)” and “food loss” from food waste sources, the methane gas generated from these raw materials, the electricity produced and CO2 reduced using methane gas, as well as the green hydrogen production volumes and associated CO2 reduction levels.
This series of programs, spanning from the IoT source investigation program (401) to the CO2 reduction IT calculation program (404), is collectively referred to as the methane resource recovery estimation system (400). User convenience is further enhanced by integrating these functions into a graphical user interface (GUI) accessible via a smartphone application.
The essential starting point for converting unused residual sludge and food loss into valuable resources is biogasification. In the context of biogasifying these unused materials, achieving optimal energy conversion equipment selection and maintaining a high utilization rate requires highly accurate methane yield predictions, as well as the prediction of inhibition and synergies among methane fermentation raw materials.
To address these needs, the methane resource recovery estimation system (400) leverages advanced AI and IoT technologies to scientifically and systematically integrate various independent data sources. By employing deep learning, the system enhances the accuracy of AI-driven predictions. Consequently, the methane resource recovery estimation system (400), as a digital transformation service in the resource recovery field, delivers new value to food waste sources and biogas plants.
Moreover, this system maximizes the utilization of food waste as a decarbonization resource, contributing significantly to the realization of carbon neutrality by 2050, both domestically and internationally.
Subsequently, the specific details of the “AI Methane Resource Recovery Program (402)” and the “AI Gasification Promotion Program (403)” executed by the evaluation value calculation device for biogas materials in this embodiment will be explained with reference to
The AI Methane Recovery Program (402) establishes a connection between food waste sources, including commercial facilities, restaurants, hotels, and food factories, and biogas (methane gas) plant operators. This connection is based on analysis results evaluating the compatibility of mixed fermentation between “biogas materials (food waste)” discharged from these sources and “biogas materials currently utilized in the plant” (e.g., food waste, oils and fats, sludge, livestock waste). Through this approach, the AI Methane Recovery Program (402) aims to enhance the efficiency of biogas power generation or clean hydrogen production, while simultaneously improving the recycling rate of food waste.
The AI Gasification Promotion Program (403) predicts and analyzes the appropriate types and quantities of gasification promoters, nutritional supplements, auxiliary raw materials, and other additives. These predictions are based on analysis and evaluation data regarding the compatibility of mixed fermentation for each biogas material, as determined by the AI Methane Resource Recovery Program (402).
Building on these insights, the AI Gasification Promotion Program (403) is designed to establish optimal fermentation conditions when one or more stored biogas materials are combined and fermented.
The assumed functions of the AI Methane Resource Recovery Program (402) and the AI Gasification Promotion Program (403) are described below. As shown in
Methane Resource Recovery Evaluation AI (402a) for Individual Materials: This AI analyzes the value of each material as a biogas resource. The analysis is based on factors such as the component properties of each material, methane yield (volume of methane gas produced), and energy yield. The evaluation covers biogas materials collected from various sources and those currently in use at each plant.
Methane Resource Recovery Evaluation AI (402b) for Mixed Materials: This AI predicts and analyzes combinations of mixed fermentation processes to identify those that achieve the highest methane yield and energy yield. The evaluation considers biogas materials collected from various sources as well as those currently utilized at each plant.
The AI Gasification Promotion Program (403) consists of the Methane Yield Promotion Prediction AI (403a), as illustrated in
Methane Yield Promotion Prediction AI (403a): This AI predicts and analyzes the types and amounts of gasification promoters required to adjust numerical indicators of analysis items to within their normal ranges. These analysis items are used to measure the value of individual materials (or their combinations) as biogas resources, as predicted and analyzed by the two AIs (402a and 402b).
Gasification promoters include auxiliary raw materials (such as oil and fat refining sludge and biodiesel fuel (BDF) residue glycerin), promoters (such as cellulase enzymes, rumen fluid, and biochar), and nutritional supplements (such as NiCl2·6H2O, CoCl2·6H2O, and FeCl2·4H2O).
All three AIs 402a, 402b, and 403a can be integrated into each device that constitutes the “methane resource matching system S” described above. These AIs provide specialized data related to resource recovery in an accessible and user-friendly format, tailored to the needs of users at both food waste sources and biogas plant operators. The data is presented with clear priorities and display formats, such as rankings based on methane yield, energy yield, CO2 emission reductions, material costs, and graphical visualizations (see
This approach addresses the previously unresolved issue of the lack of pre-diagnosis of feedstocks and the absence of component index data necessary for stable plant operation when accepting food waste as a methane feedstock. As a result, the system facilitates the promotion of biogas resource recovery from food waste.
The Methane Resource Recovery Evaluation AI 402a for individual materials is one of the AIs constituting the AI Methane Resource Recovery Program 402, as described above. This AI analyzes biogas materials collected from food waste sources-such as commercial facilities, restaurants, hotels, and food factories—as well as biogas materials currently in use at biogas plants, including food waste, oils and fats, sludge, and livestock waste.
The Methane Resource Recovery Evaluation AI 402a predicts and evaluates the value of each material as a methane resource by referencing the component index master data, as illustrated in
These results serve as reference data for compatibility analyses of mixed fermentation for each biogas material and for predicting and analyzing synergies or potential inhibition.
The assumed functions (flow) of the Methane Resource Recovery Evaluation AI 402a for individual materials are described with reference to
(A) This program facilitates the registration and validation of data into the AI Methane Resource Recovery Program 402. The data includes information on biogas materials-such as food waste collected from various sources and biogas materials currently in use at biogas plants, including food waste, oils and fats, sludge, and livestock waste—as well as data on the operational status of biogas plants.
Specific examples of data on collected food waste and biogas materials currently utilized in plants include information about the characteristics of each biogas material. This data encompasses attributes such as “collection location,” “type.” “properties” (e.g., liquid, solid, gas, processed or unprocessed), “quantity.” and inventory status. This information can be managed in real-time using technologies such as blockchain.
Additionally, data on the operating status of biogas plants encompasses various parameters. These include the type of biogas materials utilized in each plant, the presence, absence, and quantity of additives such as coagulants, the temperature of raw material storage tanks, the amount of heat required for methane fermentation tank heating, fermentation tank heating efficiency and heat loss rate, the maximum organic load per cubic meter of the methane fermentation tank, surplus heat, the amount and efficiency of hot water recovery, the organic decomposition rate within the methane fermentation tank, methane gas generation volume in the gas holder, measured methane yield (BMP), power generation amount (in the case of gas-based power generation), power generation efficiency, heat conversion values, power consumption within the facility, and the consumption ratio relative to power generation.
As with inventory data for biogas materials, this operational data can be managed in real time using technologies such as blockchain, ensuring seamless monitoring and optimization.
(B) Utilizing data on biogas materials registered and approved through the AI methane resource recovery program (402), the methane resource recovery evaluation AI (402a) analyzes and predicts the value of each material as a methane resource by referencing the component index master data.
The items verified during the prediction and analysis of the value of each material as a methane resource are assumed to include the following: the necessity of pretreatment for each methane material; organic composition; fermentation temperature; fermentation duration; organic substance concentration (VS); solid substance amount (TS); organic substance ratio (VS/TS); ash concentration; chemical oxygen demand (COD); theoretical oxygen demand (ThOD); COD/VS value; concentrations of organic constituent elements, including carbon (C: % VS), nitrogen (N: % VS), oxygen (O: % VS), and hydrogen (H: % VS); concentrations of carbohydrates, lipids, proteins, ash, and other nutrients; hemicellulose, cellulose, and lignin concentrations (% VS); C/N ratio (ratio of carbon to nitrogen concentration); lipid/VS ratio (ratio of lipid concentration to VS concentration); sodium concentration (Na); micronutrient concentrations, such as iron (mg/kg-COD feedstock), cobalt (mg/g-COD feedstock), and nickel (mg/g-COD feedstock); theoretical methane yield (ThBMP: ml/g-VS); actual methane yield (BMP: ml/g-VS) predicted value; biodegradation rate; VS decomposition rate; and energy yield.
To estimate the theoretical methane yield, multiple calculation methods are employed. These methods include those based on conventional composition formulas (elemental balance), chemical oxygen demand (COD), including both CODCr and CODMn, and nutrient-based methods that account for components such as fiber. The accuracy of the predicted and analyzed values is further enhanced by integrating data on the actual operational status of biogas plants.
For instance, when illustrating the estimation of theoretical methane yield based on a composition formula (elemental balance), the conversion of biomass to methane through methane fermentation can be represented by the following equation [Equation 1].
CaHbOcNd+(a−b/4−c/2+3d/4)H2O→(a/2+b/8−c/4−3d/8)CH4+(a/2−b/8+c/4−3d/8)CO2+dNH3 [Equation 1]
Here, CaHbOcNd represents the organic composition formula of the raw material. In practice, the organic composition is typically considered equivalent to volatile solids (VS), a widely used indicator. Volatile solids refer to the substances that are vaporized when the residue from fermentation ingredients is heated to approximately 600° C. The subscripts a to d contain values specific to the material. For raw materials with unknown composition formulas, an empirically determined formula (empirical formula) must be used, which can be obtained through methods such as the CHN Corder analysis.
For instance, reported empirical formulas include C17H29O10N or C13H21O7N for food waste, C266H434O210N for paper waste, C7H12O4N for sewage sludge, C22H31O11N for dairy cow waste, and C10H19O3N for wastewater sludge. Using these composition formulas, the theoretical biochemical methane potential (Theoretical Biochemical Methane Potential, ThBMP) is calculated as follows:
Herein, fD represents the correction coefficient, defined as the proportion of decomposable organic matter, or the fermentation decomposition rate. If the decomposition rate is unknown, it must be determined experimentally. Reported decomposition rates include 0.75 to 0.85 for food waste and food residues, 0.45 to 0.55 for pig waste, 0.25 to 0.35 for dairy cow waste, 0.34 for newspapers, 0.55 for herbaceous plants, 0.25 for pruning branches, and 0.50 for sewage sludge.
The calculation result, ThBMP, represents the volume of methane gas (in milliliters) generated per gram of volatile solids (VS) in the raw material. When employing this prediction method, it is essential to analyze the VS content of the raw material. Additionally, if the chemical formula is unknown, determining the elemental composition (CHON) of the raw material is required.
Although decomposition rates for fD have been extensively documented in various reports, when these rates are not available, they must be established through culture experiments. Such experiments involve using the raw material with methane fermentation bacteria to measure the VS decomposition rate.
To illustrate the estimation of theoretical methane yield based on nutrient composition, carbohydrates, proteins, and lipids are recognized as the three primary nutrients in food. These nutrients can be converted into methane through methane fermentation. The theoretical methane yield is calculated using [Equation 3], which incorporates the methane yields—415 ml/g VS for carbohydrates, 496 ml/g VS for proteins, and 1014 ml/g VS for lipids-calculated from the ThBMP formula in [Equation 2]. These calculations are based on the average molecular formulas C6H10O5 (carbohydrates), C5H7O2N (proteins), and C57H104O6 (lipids), as well as the weight proportions of each nutrient within the volatile solids (VS), assuming VS is normalized to a value of 1.
When employing this prediction method, it is essential to determine the VS content of the raw material and the weight ratios of the three primary nutrients—carbohydrates, proteins, and lipids. For raw materials consisting of a single food type, these values can often be obtained from the Standard Tables of Food Composition. For mixed or complex materials, analytical methods should be used to determine the proportions of carbohydrates, proteins, and lipids. Similar to the prediction method based on the elemental composition formula, the decomposition rate coefficient (fD) is derived from literature sources, such as published reports or results from culture experiments.
The component index master data, utilized by the “Methane Resource Recovery Evaluation AI 402a for Individual Materials.” includes analysis results for various sample methane materials, such as food waste, oils and fats, sludge, livestock waste, animal and plant residues, and mixed organic waste. These results encompass the following: necessity of pretreatment for each methane material; organic substance composition of each methane material; fermentation temperature; fermentation duration; organic substance concentration (volatile solids. VS); solid substance amount (total solids. TS); organic substance ratio (VS/TS); ash concentration; chemical oxygen demand (COD); theoretical oxygen demand (ThOD); COD/VS value; concentrations of organic constituent elements (e.g., carbon (C: % VS), nitrogen (N: % VS), oxygen (O: % VS), hydrogen (H: % VS)); nutrient concentrations (e.g., carbohydrates, lipids, proteins, ash); hemicellulose, cellulose, and lignin concentrations (% VS); C/N ratio (carbon-to-nitrogen concentration ratio); lipid/VS ratio (lipid concentration to VS concentration ratio); sodium (Na) concentration; micronutrient concentrations (e.g., iron (mg/kg-COD feedstock), cobalt (mg/g-COD feedstock), nickel (mg/g-COD feedstock)); theoretical methane yield (ThBMP: ml/g-VS); actual methane yield (BMP: ml/g-VS); biodegradation rate; VS decomposition rate; energy yield; and more.
In addition to the analysis results for each of these items, data on the actual operating conditions at individual biogas plants is sampled for each methane material utilized and incorporated as learning data for the Methane Resource Recovery Evaluation AI 402a for Individual Materials.
(C) The prediction and analysis results generated by the Methane Resource Recovery Evaluation AI 402a for Individual Materials are presented within the Methane Resource Matching System to both the distributors of each material and the biogas plant operators. Simultaneously, these prediction and analysis results are imported into subsequent processes, including the “Methane Resource Recovery Evaluation AI 402b for Mixed Materials” and the “AI Gasification Promotion Program 403.” These data serve as reference points for predicting and analyzing the compatibility of mixed fermentation for various biogas materials, as well as their synergistic (or inhibitory) effects.
The Methane Resource Recovery Evaluation AI 402b for Mixed Materials is one of the core components of the “AI Methane Resource Recovery Program 402.” This AI performs compatibility analysis and evaluation for mixed fermentation of one or more stocked biogas materials. It relies on the evaluation, prediction, and analysis data related to methane resource recovery generated by the aforementioned “Methane Resource Recovery Evaluation AI 402a for Individual Materials.”
The evaluation considers biogas materials collected from various sources (e.g., commercial facilities, restaurants, hotels, food factories) and biogas materials currently used in biogas plants, including food waste, oils and fats, sludge, and livestock waste. By referring to the mixed index master data shown in
By simultaneously referencing data on the operational status of each biogas plant, the Methane Resource Recovery Evaluation AI 402b for Mixed Materials can analyze and identify the biogas material most compatible for mixed fermentation at each biogas plant. Furthermore, it can determine, in real time, the distribution source capable of providing the optimal biogas material for mixed fermentation.
In this context, data related to each biomass material, its source of origin, the biogas plant, and the biogas materials currently in use at the plant are always interconnected and maintained as a unified dataset.
The operational flow for the Methane Resource Recovery Evaluation AI 402b for Mixed Materials, as illustrated in
(E) Building on the prediction and analysis results for individual biogas materials imported from the Methane Resource Recovery Evaluation AI 402a, the Methane Resource Recovery Evaluation AI 402b for Mixed Materials evaluates the compatibility of methane fermentation for mixed biogas materials. This is achieved by referencing the Mixture Index Master Data to predict and analyze the effects of combining and fermenting the materials.
The following items are assumed as prediction and analysis parameters when evaluating the compatibility of mixed fermentation for stocked biogas materials: the necessity of pretreatment for each raw material during mixing and fermentation; the organic composition of each methane material; fermentation temperature and days; concentrations of organic substances (VS) and solids (TS); the VS/TS ratio; ash concentration; chemical oxygen demand (COD); theoretical oxygen demand (ThOD); the COD/VS ratio; elemental composition of organic substances, including carbon (C: % VS), nitrogen (N: % VS), oxygen (O: % VS), and hydrogen (H: % VS); concentrations of nutrients such as carbohydrates, lipids, proteins, and ash; hemicellulose, cellulose, and lignin concentrations (% VS); the C/N ratio (carbon-to-nitrogen concentration ratio); the lipid/VS ratio; Na concentration; micronutrient concentrations such as iron (mg/kg-COD raw material), cobalt (mg/g-COD raw material), and nickel (mg/g-COD raw material); theoretical methane yield (ThBMP: ml/g-VS); actual methane yield (BMP: ml/g-VS); biodegradation rate; VS decomposition rate; and energy yield and more.
To estimate the theoretical methane yield, multiple calculation methods are employed. These methods include composition formula-based calculations (elemental balance), chemical oxygen demand (COD)-based calculations (including CODCr and CODMn), and nutrient-based calculations (which incorporate fiber content). Additionally, the accuracy of prediction and analysis values is enhanced by integrating data on the actual operating conditions of biogas plants.
The mixed index master data, which the “methane resource recovery evaluation AI 402b for mixed materials” utilizes for learning and analysis, includes data for various sample methane materials, such as food waste, oils and fats, sludge, livestock waste, animal and plant residues, and mixed organic waste. The data encompasses the following attributes: whether pretreatment of each methane material is necessary; the organic substance composition of each methane material; fermentation temperature; fermentation days: organic substance concentration (VS); solid substance amount (TS); organic substance ratio (VS/TS); ash concentration; chemical oxygen demand (COD); theoretical oxygen demand (ThOD); COD/VS value; organic substance constituent element concentrations, such as carbon (C: % VS), nitrogen (N: % VS), oxygen (O: % VS), and hydrogen (H: % VS); carbohydrate, lipid, protein, ash, and other nutrient concentrations; hemicellulose, cellulose, and lignin concentrations (% VS); the C/N ratio (carbon-to-nitrogen concentration ratio); lipid/VS ratio (lipid concentration to VS concentration ratio); Na concentration; micronutrient concentrations such as iron (mg/kg-COD raw material), cobalt (mg/g-COD raw material), and nickel (mg/g-COD raw material); theoretical methane yield (ThBMP: ml/g-VS); actual methane yield (BMP: ml/g-VS) predicted value; biodegradation rate; VS decomposition rate; and energy yield.
The analysis results for each of these attributes are incorporated into the master data. Additionally, data on the actual operating conditions at biogas plants is sampled based on the methane materials used and is employed as learning data for the AI.
The evaluation value calculation device described in this embodiment utilizes specific indices and logic to determine the compatibility of each biogas material in stock during mixed fermentation. For example, the following items, in the order a through g, are assumed: (a) C/N ratio: whether the ratio of carbon concentration to nitrogen concentration in the raw material is 25 or more; values below 25 indicate a risk of inhibition;
Additionally, if the methane yield is less than 100 mL/g-VS, the combination of biogas materials is considered unsuitable for simple mixed fermentation from an economic perspective.
Furthermore, the prediction and analysis results (data) generated by the methane resource recovery evaluation AI 402b for mixed materials are presented to both the sources of the materials and biogas plant operators through the methane resource matching system S. Simultaneously, these results are imported into the subsequent “AI Gasification Promotion Program 403” to serve as reference data for predicting and analyzing the synergy and inhibitory effects of each biogas material during mixed fermentation.
(F) The results (data) of the predictions and analyses conducted by the methane resource recovery evaluation AI 402b for mixed materials are presented to the sources of each material and to biogas plant operators via the methane resource matching system S. Simultaneously, these prediction and analysis results are imported into the subsequent “AI Gasification Promotion Program 403” and serve as one of the reference data sets for predicting and analyzing the synergy and inhibitory effects of each biogas material during mixed fermentation.
The AI gasification promotion program 403 comprises a methane yield promotion prediction AI 403a; it forms part of the methane resource matching system S, together with the previously mentioned AI methane resource recovery program 402.
The methane yield promotion prediction AI 403a predicts and analyzes the types and amounts of “gasification promoters.” “nutritional supplements,” and “auxiliary raw materials” required to create optimal fermentation conditions when mixing and fermenting one or more biogas materials in stock. These predictions are based on evaluation and analysis data from the “Methane resource recovery evaluation AI 402b for mixed materials,” which examines biogas materials collected from sources such as commercial facilities, restaurants, hotels, and food factories, alongside materials currently used in biogas plants, including food waste, oils and fats, sludge, and livestock waste. Furthermore, compatibility analysis and evaluation data for mixed fermentation of in-stock biogas materials, also provided by the AI, are incorporated into this process. The analysis is further refined by referencing the bio-gasification promotion index master data illustrated in
The methane yield promotion prediction AI 403a predicts and analyzes the type and amount of gasification promoters required to adjust the numerical indicators of each analysis item that influence methane gas production. These adjustments ensure that items with insufficient or excessive values are brought within their normal ranges. The indicators include, but are not limited to, the conditions a to g listed for determining the compatibility of the aforementioned mixed fermentation.
The methane yield promotion prediction AI 403a predicts and analyzes the potential synergy associated with the mixing of materials that do not require the addition of gasification promoters or other additives (i.e., those that satisfy all numerical indicators for biogas generation). This analysis is conducted by referencing the bio-gasification promotion indicator master data, as shown in
Additionally, for combinations of materials identified through the prediction and analysis results of the “Methane resource recovery evaluation AI 402b for mixed materials” as having an estimated methane yield of less than 100 mL/g-VS after mixed fermentation, the addition of “gasification promoters.” “nutritional supplements,” or “auxiliary raw materials” may enable the combination to meet favorable conditions for methane gas production.
The functionality and process flow assumed for the methane yield promotion prediction AI 403a are described below with reference to the flow diagram presented in
(G) Initially, the evaluation, prediction, and analysis data generated by the “methane resource recovery evaluation AI 402a for individual materials” and the “methane resource recovery evaluation AI 402b for mixed materials” are integrated into the AI gasification promotion program 403 (methane yield promotion prediction AI 403a) through system integration.
(H) Based on the prediction and analysis data imported from the “Methane Resource Recovery Evaluation AI 402a for Individual Materials” and the “Methane Resource Recovery Evaluation AI 402b for Mixed Materials.” the Methane Yield Promotion Prediction AI 403a predicts and analyzes the types and quantities of “gasification promoters.” “nutritional supplements.” and “auxiliary raw materials” necessary to optimize fermentation conditions for the individual fermentation of each biogas material. This process is carried out by referencing the bio-gasification promotion index master data.
Furthermore, the Methane Yield Promotion Prediction AI 403a evaluates the types and quantities of “gasification promoters.” “nutritional supplements,” and “auxiliary raw materials” required to optimize fermentation conditions for various combinations of one or more stocked biogas materials. These combinations are derived from the “Methane Resource Recovery Evaluation AI 402b for Mixed Materials.” and the analysis is also conducted by referencing the bio-gasification promotion index master data.
Herein, the types of “gasification promoters.” “nutritional supplements.” and “auxiliary raw materials” considered include the following examples: gasification promoters, such as cellulase enzymes, rumen fluid, and biochar; nutritional supplements, such as NiCl2·6H2O, CoCl2·6H2O, and FeCl2·4H2O; and auxiliary raw materials, such as oil and fat refining sludge and BDF residue glycerin.
For example, during the individual or mixed fermentation of each biogas material, the methane yield promotion prediction AI 403a predicts and analyzes the types and quantities of “gasification promoters.” “nutritional supplements.” and “auxiliary raw materials” required to ensure that the values of the following analytical items fall within the optimal range for favorable fermentation conditions: whether pretreatment of raw materials is necessary; the organic composition of each methane material; fermentation temperature; fermentation duration; organic substance concentration (VS); solid substance amount (TS); organic substance ratio (VS/TS); ash concentration; chemical oxygen demand (COD); theoretical oxygen demand (ThOD); COD/VS value; concentrations of organic elemental constituents in each methane material, including carbon (C: % VS), nitrogen (N: % VS), oxygen (O: % VS), and hydrogen (H: % VS); carbohydrate, lipid, protein, ash, and other nutrient concentrations; hemicellulose, cellulose, and lignin concentrations (% VS); C/N ratio (carbon-to-nitrogen concentration ratio); lipid/VS ratio (lipid-to-VS concentration ratio); Na concentration; concentrations of micronutrients such as iron (mg/kg-COD raw material), cobalt (mg/g-COD raw material), and nickel (mg/g-COD raw material) during material mixing; theoretical methane yield (ThBMP: ml/g-VS); predicted actual methane yield (BMP: ml/g-VS); biodegradation rate; VS decomposition rate; and energy yield.
The numerical indicators for each analysis item include the following criteria:
In addition to the primary indicators, further analysis items and numerical indicators used to assess synergy include the following criteria:
For individual materials (data from the methane resource recovery evaluation AI 402a), the synergy resulting from mixing the material with a “gasification promoter” or similar additives is predicted and analyzed. For mixed materials (data from the methane resource recovery evaluation AI 402b), the synergy from simple mixing of materials, both with and without the addition of “gasification promoters” or similar additives, is also predicted and analyzed.
(I) Finally, the methane yield promotion prediction AI 403a outputs the results of its predictions and analyses, which are then presented to the sources of the respective biogas materials and the operators of biogas plants through the methane resource matching system S.
The processing units included in the evaluation value calculation device 1 for biogas materials, as implemented in this embodiment, are described with reference to
First, the processing units integrated within the methane resource recovery device 1 will be described with reference to
The control unit (10) utilizes a processor, such as a CPU, and memory to manage the device's components and execute various functions.
The evaluation value calculation unit (11) employs a processor and memory to estimate evaluation values related to the resource recovery of the target biogas material, based on control instructions received from the control unit (10). The control unit (10) and the evaluation value calculation unit (11) may be implemented as a single hardware unit, such as a System on a Chip (SoC), integrating a processor (e.g., a CPU), memory, the storage unit (12), and the communication unit (13).
The storage unit (12) utilizes a hard disk or flash memory to store various essential components, including the evaluation value calculation program (1P) and a machine learning library (1L) that serves as the machine learning model. Additionally, the storage unit (12) retains definition data for the machine learning model, parameters such as the trained model's setting values, and other related information. The programs contained within the evaluation value calculation program (1P) specifically include the AI methane resource recovery program (402) and the AI gasification promotion program (403) previously described.
The communication unit (13) functions as a communication module enabling connectivity to communication networks, such as the Internet. It facilitates the transmission of biogas material-related information, including a machine learning library and evaluation values associated with the resource recovery of biogas materials, to other terminal devices. The communication unit (13) may incorporate components such as a network card, a wireless communication device, or a carrier communication module.
The display unit (14) utilizes components such as a liquid crystal panel to provide visual output. The operation unit (15) may include user interface devices such as a keyboard or mouse. It transmits user-input operation information to the control unit (10). This user-provided information may include property data (e.g., type, collection location, characteristics, and quantity of biogas materials, both individual and mixed) as well as operational status data of the biogas plant.
The reading unit (16) may utilize a disk drive or similar device to access the optimization calculation program and machine learning library stored on a recording medium (e.g., an optical disk). The evaluation value calculation program (1P) and machine learning library (1L) stored in the storage unit (12) can originate from the evaluation value calculation program (2P) and machine learning library (3L) read by the reading unit (16) from the recording medium. These programs and libraries may then be copied to the storage unit (12) under the control of the control unit (10).
Next, the functionality of the evaluation value calculation device (1) is described with reference to
The learning process execution unit (101) functions as a machine learning model (machine learning engine) based on the machine learning library (1L), definition data, and parameter information stored in the storage unit (12). In essence, the learning process execution unit (101) executes the process of automatically learning the setting values (e.g., parameters) of the machine learning model to be trained, using training data while utilizing the model itself during the learning process.
During the learning process, the machine learning libraries analyze various parameters related to biogas materials. These parameters include whether pretreatment is required for different biogas materials; the organic substance composition of each methane material; fermentation conditions such as temperature and days; organic substance concentration (VS: Volatile Solids) and solid substance amount (TS: Total Solids); the ratio of organic substances (VS/TS); ash concentration; chemical oxygen demand (COD); theoretical oxygen demand (ThOD); COD/VS value; the concentration of organic constituent elements (e.g., carbon (C: % VS), nitrogen (N: % VS), oxygen (O: % VS), hydrogen (H: % VS)); the concentrations of carbohydrates, lipids, proteins, ash, and other nutrients; hemicellulose, cellulose, and lignin concentrations (% VS); the C/N ratio (carbon to nitrogen concentration ratio); lipid/VS ratio (lipid to VS concentration ratio); Na concentration; and micronutrient concentrations such as iron (mg/kg-COD raw material), cobalt (mg/g-COD raw material), and nickel (mg/g-COD raw material). Additional factors include theoretical methane yield (ThBMP: ml/g-VS), predicted actual methane yield (BMP: ml/g-VS), biodegradation rate, VS decomposition rate, and energy yield.
For example, the learning process execution unit 101 can perform a process to minimize the error between the output data generated by inputting training data into the complete machine learning model 112 and the corresponding known training data. As a result, the unit updates the model's parameters (weights). The parameters obtained through this learning process are stored in the storage unit 12.
The evaluation value calculation execution unit 102 estimates evaluation values based on the evaluation value calculation program 1P stored in the storage unit 12. Specifically, it uses a machine learning model to estimate evaluation values for the input biogas materials (individual or mixed). Additionally, the evaluation value calculation execution unit 102 processes input data, which includes biogas material-related information (e.g., machine learning libraries, property data such as the type of biomass material, collection location, characteristics, quantity, and operational status data of the biogas plant). This input is provided to the input unit 111 based on user operations performed through the operation unit 15.
As described above, the operational status data for the biogas plant includes information such as the type of biogas material used at the plant; the fermentation method; the presence and quantity of additives, such as coagulants; the volatile solids (VS) volume load per cubic meter of the methane fermentation tank; the organic decomposition rate within the methane fermentation tank; the volume of methane gas generated in the gas holder; and the measured methane yield (Biochemical Methane Potential, BMP).
The input unit 111 of the evaluation value calculation unit 11 processes input related to a machine learning library associated with biogas materials, as well as biogas material-related data. The learning process execution unit 101 transmits the machine learning library to the machine learning model 112. Additionally, the evaluation value calculation execution unit 102 provides property data of the biogas material, which is the subject of the evaluation value estimation, along with operational status data of the biogas plant, to the machine learning model 112.
The specific type of biogas material-related data, including property data of the biogas material under evaluation and operational status data of the biogas plant, may be identified, for example, through information or an identifier located in the header section.
The development program for the machine learning model 112 may utilize widely adopted programming languages such as Python or MATLAB. These tools are commonly employed for tasks such as reading various data files, performing deep learning, conducting data analysis, constructing model formulas, and implementing programming logic.
If a convolutional neural network (CNN) is employed, which utilizes multiple convolutional layers and pooling layers to extract features from input data during inference, the machine learning model 112 may incorporate several stages of convolutional and pooling layers as defined by the corresponding definition data. Additionally, the model may include a fully connected layer to extract input data features and execute estimation processing based on these extracted features.
When utilizing a trained model, the machine learning model 112 estimates evaluation values related to the resource recovery of biogas materials, based on pre-trained parameters. These evaluations are derived from property data, such as the type, collection location, characteristics, and quantity of biogas materials (both individual and mixed materials), as well as data on the operational status of the biogas plant. The evaluation values include determinations such as whether pretreatment is required; the organic substance composition of each methane material; fermentation temperature; fermentation duration; organic substance concentration (VS); solid substance amount (TS); organic substance ratio (VS/TS); ash concentration; chemical oxygen demand (COD); theoretical oxygen demand (ThOD); COD/VS ratio; concentrations of organic substance constituent elements (e.g., carbon (C: % VS), nitrogen (N: % VS), oxygen (O: % VS), hydrogen (H: % VS)); concentrations of nutrients such as carbohydrates, lipids, proteins, and ash; hemicellulose, cellulose, and lignin concentrations (% VS); C/N ratio (carbon to nitrogen concentration ratio); lipid/VS ratio (lipid to VS concentration ratio); sodium (Na) concentration; and micronutrient concentrations such as iron (mg/kg-COD feedstock), cobalt (mg/g-COD feedstock), and nickel (mg/g-COD feedstock). Additionally, theoretical methane yield (ThBMP: ml/g-VS), actual methane yield (BMP: ml/g-VS) predicted values, biodegradation rate. VS decomposition rate, energy yield, and other related metrics are estimated.
Particularly, when estimating the evaluation value of mixed materials, the following criteria are assessed:
If the methane yield is less than 100 mL/g-VS, the combination of materials is deemed unsuitable for simple mixed fermentation from an economic perspective.
Herein, the machine learning model 112 assesses the appropriate types and quantities of gasification promoters, nutritional supplements, auxiliary raw materials, and related components, based on the evaluation values of the mixed fermentation of each biogas material. Consequently, it becomes possible to optimize fermentation conditions for the processing of one or more stored biogas materials.
In this context, the evaluation values used to determine the synergy with the “gasification promoter,” “nutritional supplement,” and “auxiliary raw material” are specified as follows:
The evaluation value generated by the machine learning model 112 is subsequently sent to the output unit 113 and can be stored in the memory unit 12 as biogas material-related data.
As described above, the evaluation value calculation device 1 links biogas materials discharged from their sources (e.g., commercial facilities, restaurants, hotels, food factories) with the biogas materials currently utilized in biogas plants, using the evaluation values. This linkage enables the preparation of a machine learning model designed for more efficient biogas production. Consequently, the evaluation value calculation device 1 employs the AI-driven machine learning model 112 to learn and estimate evaluation values related to the resource recovery of biogas materials. This capability facilitates precise mixing of methane gas materials and contributes to achieving carbon neutrality.
In the next step, the operation procedure for the evaluation value calculation device 1, functioning as a learning device in this embodiment, will be described with reference to
The operation procedure for the evaluation value calculation device 1, when functioning as an estimation device in this embodiment, is described below with reference to
The following section presents an experimental example of machine learning using the evaluation value calculation device 1 for methane materials, as described in this embodiment, with reference to
The AI model developed in this experiment is designed to predict evaluation values such as methane yield in biogas materials. For the purpose of this experiment, “MATLAB” (produced by Math Works) was selected as the AI model development platform due to its robust mathematical processing capabilities, integration of advanced neural networks, extensive machine learning functions, and its ability to operate without requiring vast amounts of training data or complex programming. Alternative platforms, such as “Python” or “Mathematica” (produced by Wolfram), may also be utilized for similar purposes.
The specific AI model developed in this experiment is described below. An example of the machine learning library utilized in this AI model is presented in Tables 140 and 150, shown in
It is possible to include various types of AI models in this context, such as a machine learning model that trains a high-dimensional dataset as a multivariate normal distribution, and a deep learning model utilizing an autoencoder, which is a specific type of deep learning. However, other learning models may also be implemented, depending on the requirements and available data.
The estimated results of the evaluation values for resource recovery in this experimental example are described with reference to
In Table 181 of
Using the created machine learning model, evaluation values for each mixed biogas material were estimated. In this scenario, one material (potato waste 161) from the individually estimated biogas materials was mixed at a 50:50 ratio with each of the other individual biogas materials listed in
In this analysis, the mixed biogas materials that meet the minimum required conditions (a, b, e, and f) are listed above the dashed line in the tables, with combinations arranged in descending order of methane yield. Materials listed below the dashed line 191 are considered inappropriate for mixing, even if they result in a higher methane yield, because they fail to meet one or more of the required conditions (a, b, e, or f). Specifically, as illustrated in
Furthermore, even if certain mixed biogas materials do not initially satisfy all of the required conditions (a to f), it is possible for them to meet these conditions with the addition of auxiliary agents such as “gasification promoters.” “nutritional supplements.” or “auxiliary raw materials.” To evaluate this possibility, further estimations were performed by adding a gasification promoter (bioash, in this case, at a ratio of 2 kg per 100 kg of raw material) to the previously estimated evaluation values of the mixed biogas materials shown in
In this case, the mixed biogas materials that meet all the conditions (a to f) are displayed above the dashed line, and the combinations yielding the highest methane production are listed in descending order. It was determined that the materials listed below dashed line 221 do not form a suitable mixture, as they fail to satisfy any of the conditions a through f, even if the methane yield is high.
As illustrated in
Furthermore, it was estimated that the methane yield would exceed the previously estimated value when no gasification promoter was used (methane yield without gasification promoter: 420.33; methane yield with gasification promoter: 622.17). Based on this analysis, for the individual biogas material “potato waste.” the optimal condition for enhanced biogas production was determined to involve mixing it with rice stalk-BP and incorporating a gasification promoter.
As described above, this disclosure pertains to an evaluation value calculation device (Device 1) for biogas materials that calculates evaluation values related to resource recovery when biogas is produced using at least one or more types of biogas materials. The device includes the following components: an input unit (111), which receives input data including a machine learning library (1L) related to biogas materials, property data concerning biogas materials, or operational status data from a biogas plant; a machine learning model (112), designed to accept inputs from the input unit, such as the machine learning library (1L), property data, or operational status data, and to utilize this information for learning and estimation processes; a learning process execution unit (101), which trains and adjusts the parameters of the machine learning model (112) based on the machine learning library (1L); and an evaluation value calculation execution unit (102), which utilizes the machine learning model (112) to estimate evaluation values related to the resource recovery of biogas materials based on property data or operational status data.
The property data includes at least one of the following: type; collection location; properties; or quantity of biomass materials.
The evaluation values include at least one of the following indicators related to mixed biogas materials when individual biogas materials or combinations thereof are mixed: (a) C/N ratio (the ratio of carbon concentration to nitrogen concentration); (b) Lipid/VS ratio (the ratio of lipid concentration to volatile solids concentration); (c) Lipid/Carbohydrate ratio; (d) Protein/Carbohydrate ratio; (e) Micronutrient concentrations, such as iron (mg/kg-COD feedstock), cobalt (mg/g-COD feedstock), and nickel (mg/g-COD feedstock); (f) Sodium (Na) concentration; (g) Theoretical methane yield (ThBMP: ml/g-VS).
With this configuration, the evaluation value calculation device for biogas materials enables the active utilization of biogas resources by providing specific evaluation values. This disclosure addresses key challenges, such as the lack of prior diagnostics for biogas materials and the absence of reliable evaluation values for ensuring stable operation of biogas plants-both of which have hindered the use of food waste and other materials as biogas feedstocks.
Furthermore, by leveraging the estimated evaluation values, the disclosure facilitates optimal matching of food waste, residual sludge, food loss, and similar resources with biogas plants that perform methane gasification. Through the methane resource matching system accessible via the Internet, the disclosure enhances the operating rate of biogas plants and contributes significantly to achieving a decarbonized society.
With this configuration, the evaluation value calculation device 1 for biogas materials facilitates the more effective utilization of biogas resources by providing specific and actionable evaluation values. Consequently, this disclosure addresses key challenges, including the absence of prior diagnostic tools for biogas materials and the lack of reliable evaluation metrics to ensure the stable operation of power plants that utilize such materials. These challenges have historically hindered the adoption of food waste and similar materials as viable sources for biogas production.
Moreover, leveraging the estimated evaluation values, the disclosure enables the optimal pairing of food sources: such as residual sludge; food loss; and other waste materials, with biogas plants that employ methane gasification processes for food waste. This matching process is facilitated through the methane resource matching system, accessible via the Internet. By improving the operational efficiency of biogas plants, this innovation contributes significantly to advancing the transition toward a decarbonized society.
It is important to note that to achieve the objectives of the evaluation value calculation device, the present disclosure may also be implemented as a determination method comprising steps corresponding to the functional units included in the evaluation value calculation device. Additionally, the disclosure can be embodied as a program that enables a computer to execute these steps. This program may be distributed via a physical recording medium, such as a USB drive, or through a transmission medium, such as the Internet.
This application is a national phase application under 35 U.S.C. 371 of International Application No. PCT/JP2023/004304 filed Feb. 9, 2023, entitled “BIOGAS MATERIAL EVALUATION VALUE COMPUTATION DEVICE AND PROGRAM USED FOR BIOGAS MATERIAL EVALUATION VALUE COMPUTATION DEVICE,” the contents of which being incorporated by reference in its entirety herein.
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/JP2023/004304 | 2/9/2023 | WO |
