Digital twin-based renewable fuel plant systems and operations

Abstract

The present invention provides a digital twin-based renewable fuel plant system comprising: first data which is measured data; a database that stores the first data in large capacity at high speed; an AI modeling engine that performs an algorithm based on the first data stored in the database to provide second data; an intelligent central control system that receives the second data modeled by the AI ​​modeling engine; a 3D visualization engine that visualizes the received data in 3D in the intelligent central control system; and a DT computation engine that performs computations on one or more of the database, the intelligent central control system, and the 3D visualization engine.

Description

Technology Field

[0001] The present invention relates to a digital twin-based renewable fuel plant execution device and operation method, and more specifically, to a system that selects from among prediction algorithms and generates virtual data for modeling and simulation. Background Technology

[0003] A digital twin is a virtual model that describes dynamic properties while maintaining identity and consistency with a real-world physical system (e.g., transportation systems, production systems, etc.) <e.g., discrete event models, physics-based models, statistical models, machine learning models, etc.>.

[0004] By performing simulations on a digital twin developed using the operating procedures / rules of a real-world system of interest, real-time operational data, aggregated data of the same group, operational history, maintenance history, Failure Mode and Effects Analysis (FMEA), Finite Element Analysis (FEA) models, CAD models, etc., according to scenarios formed by combinations of input variables, it is possible to analyze the behavior of the actual system, diagnose failures, or achieve operation-related prediction and optimization.

[0005] Meanwhile, modeling simulation technology, which enables simulation in a virtual world based on scenarios formed from input variable values ​​for system analysis through various virtual experiments that are difficult or impossible to perform in the real world due to reasons such as excessive costs for analysis when using actual systems, safety risks during execution, and the inability to apply them to objects that do not exist in reality but only exist as concepts or plans, has long been widely used as a scientific analysis method in almost all domains.

[0006] Bioenergy, which is recently evaluated as one of the alternative energy sources to address issues such as the depletion of fossil fuels, safety concerns regarding nuclear power generation, and the high costs and low efficiency of using other energy sources, can be used in various ways depending on its purpose and form, including biogas, bioethanol, biodiesel, methanol, and hydrogen, as well as for heating, automotive fuel, and power generation fuel.

[0007] Biomass refers to organisms, such as plants and microorganisms, that can be used as energy sources for bioenergy and are produced during ecosystem cycles.

[0008] Representative biomass includes woody materials such as grains, waste wood, plant stems, and tree bark, as well as seaweed, animal manure, food waste, and organic wastewater. These materials are used to extract bioenergy through pyrolysis or fermentation processes, or converted into products such as compost, animal feed, and various plastics.

[0009] Biomass fuels generally exhibit characteristics such as heterogeneity in composition and content, and water reabsorption due to low density and calorific value, high moisture content, difficulties in crushing and grinding, and various types and growth environments. As a result, it is difficult to produce synthesis gas with a stable composition, so thermal pretreatment such as torrefaction must be performed beforehand.

[0010] In conventional biomass gasification, a large amount of energy is required for the moisture evaporation, heating of combustible matter, evaporation of volatile matter, and heating of ash during the moisture evaporation and devolatilization processes when biomass is input. In particular, since the energy required for moisture evaporation is the highest, moisture content has an absolute impact, so moisture content control must be performed through pretreatment.

[0011] Furthermore, while quality can be improved through the aforementioned preprocessing, conventionally, operations have been performed based on the operator's experience, leading to problems such as inconsistent quality or the production of low-quality products. Additionally, the absence of a system utilizing digital twins has resulted in issues such as energy waste.

[0012] Korean Published Patent Application No. 2022-0081880 relates to a device for real-time analysis and real-time operation visualization of a discrete event system using a digital twin, and is manufactured for a discrete event system, and a digital twin (110) that generates big data to generate a performance index machine learning data model (121) for outputting an analysis performance index of the discrete event system by performing a simulation in a virtual world according to an experiment scenario formed with input variable values ​​for analyzing the system through various virtual experiments, and a visualization log file machine learning data model (122) for generating a visualization log file for visualizing the operation of the discrete event system; The present invention provides a discrete event system real-time analysis and real-time operation visualization device using a digital twin, characterized by comprising: a data model generation and execution unit (120) that generates the performance index machine learning data model (121) and the visualization log file machine learning data model (122) by machine learning big data obtained using the simulation of the digital twin (100); executes the performance index machine learning data model (121) with a problem scenario formed as an input variable value for changing the operation scenario of the discrete event system during the analysis process for evaluating the performance of the discrete event system, outputs an analysis performance index of the discrete event system reflecting the problem scenario, and executes the visualization log file machine learning data model (122) with the problem scenario to generate a visualization log file reflecting the problem scenario.

[0013] Korean Registered Patent Publication No. 10-2488667 relates to a solar power generation prediction system using machine learning, comprising: a server that collects climate data and power generation data in real time from a historical power generation database, a historical climate database, a current power generation database, and a current climate database; a solar radiation and temperature analysis modeling unit that calculates module temperature data using ambient temperature and wind speed, and the solar radiation on the array inclined surface when reaching the solar module by utilizing the sun's position and the tilt angle and azimuth angle of the solar module; a first machine learning unit that generates a prediction model by setting historical power generation data, climate data, array inclined surface solar radiation, and module temperature data from the server as input values, finding the correlation between climate data and power generation using machine learning, learning, testing, and verifying, and then testing and verifying the prediction model using current climate and power generation data; and a first power generation prediction unit that predicts the power generation of a solar module array from the first machine learning unit. A second machine learning unit that predicts the solar radiation distribution of a corresponding power plant by setting power plant system information, solar radiation data, and cloud information as input values, analyzes cloud characteristics through solar radiation data, finds correlations between cloud characteristics and cloud information to create a prediction model, and then trains, tests, and verifies it; a cloud loss calculation module that calculates solar radiation loss due to clouds in a solar power plant from the second machine learning unit; and a second power generation prediction unit that derives an optimal power generation amount as a result through the predicted power generation amount from the first power generation prediction unit and the power generation loss amount according to the cloud loss calculation module, wherein the solar radiation and temperature analysis modeling unit includes a location data unit that collects longitude and latitude data of a prediction point on a specific date from a server; and a solar position calculation unit that calculates the solar position according to the date, longitude, and latitude collected by the location data unit.A solar power generation prediction system using machine learning is provided, characterized by comprising: a solar irradiance calculation unit that calculates the solar irradiance of the array tilt surface (POA) irradiated on the solar power generation module by utilizing possible azimuth and tilt angles for the location of the solar power plant with respect to the solar azimuth and altitude obtained from the solar position calculation unit; and a temperature calculation unit that calculates the temperature of the solar power generation module through climate data of the solar power generation module, ambient temperature, and wind speed.

[0014] Korean Published Patent Application No. 2022-0097252 relates to a machine learning-based equipment management method and system for a smart plant. The equipment management method performed by a machine learning-based equipment management system for a smart plant comprises: (a) a step of collecting key parameter information for a plurality of pipes used in the plant; (b) a step of clustering the plurality of pipes into a first group to an Nth group through a clustering algorithm based on the collected key parameter information; and (c) a step of automatically setting an inspection cycle for each group based on the results of thickness measurement inspections of pipes stored for each group or historical information based on the frequency of work due to the occurrence of abnormal phenomena.

[0015] Korean Registered Patent Publication No. 10-2601707 relates to a digital twin system and method capable of real-time monitoring, virtual operation prediction, and optimized operation services for field equipment, comprising: field equipment equipped with sensors; a digital twin construction service platform that generates digital twin data for said field equipment; and a user terminal on which digital twin data downloaded from said digital twin construction service platform is loaded; wherein the user terminal comprises a sensor data interface that receives field data including operating conditions and measurement data transmitted in real time from said sensors; A digital twin application comprising: a result prediction module that monitors field equipment or predicts results regarding virtual operating conditions of field equipment by linking the mounted digital twin data with field data input through the sensor data interface; and a data fusion module that updates the prediction performance of the mounted digital twin data for the entire operating range by utilizing operating conditions and measurement data received through the sensor data interface; wherein the digital twin data includes a principal component vector forming an order reduction model and a function that predicts principal component coefficients for arbitrary operating conditions; and wherein the data fusion module obtains principal component coefficients corresponding to measurement data through a data fusion technique, collects data regarding the obtained principal component coefficients and operating conditions for a certain period of time over the entire operating range, and updates the function that predicts principal component coefficients for operating conditions through machine learning on the collected data.

[0016] Japanese Patent Publication No. 2022-537799 relates to a method and means for optimizing biotechnological production, and provides a novel method for the automatic generation and verification of a digital twin for the production of biotechnological products and the application of the digital twin for the purpose of increasing product concentration, productivity, biomass concentration, and product quality by optimizing the media composition and / or feed profile.

[0017] However, in conventional combustion devices for drying, carbonization, and backcarbonization, condition control was based on the operator's experience, which resulted in problems such as low product quality, compromised consistency, and irregular performance.

[0018] Furthermore, digital twin technology applied to conventional combustion devices and equipment, such as drying, carbonization, and torrefaction, has not been disclosed.

[0019] Furthermore, unnecessary energy was being wasted by operating the feed amount and speed in a fixed manner.

[0020] Therefore, no technology has been disclosed for reducing energy waste through a digital twin method in combustion devices such as drying, carbonization, and torrefaction, and for solving conventional problems through modeling and simulation. Prior art literature

[0022] Korean Published Patent Application No. 2022-0081880 is Korean Registered Patent Application No. 10-2488667, Korean Published Patent Application No. 2022-0097252, Korean Registered Patent Application No. 10-2601707, and Japanese Published Patent Application No. 2022-537799 The problem to be solved

[0023] The present invention aims to solve the problems of the prior art and provide a system that generates virtual data for modeling and simulation by selecting a prediction algorithm. means of solving the problem

[0025] To solve the above problems, the present invention provides a digital twin-based renewable fuel plant system comprising: first data which is measured data; a database that stores the first data in large capacity at high speed; an AI modeling engine that performs an algorithm based on the first data stored in the database to provide second data; an intelligent central control system that receives the second data modeled by the AI ​​modeling engine; a 3D visualization engine that visualizes the received data in 3D in the intelligent central control system; and a DT computation engine that performs computations on one or more of the database, the intelligent central control system, and the 3D visualization engine.

[0026] In addition, the first data may be any one or more of the following: primary combustion furnace TC, secondary combustion furnace TC, combustion furnace outlet temperature, B / F temperature, differential pressure, pulsing history, pellet input amount, calorific value, ternary, elemental adjustment, combustion furnace temperature, NOx, SOx, O2, and combustion chamber flame image.

[0027] In addition, the above AI modeling engine may include a data statistical prediction model.

[0028] In addition, the above intelligent central control system can perform one or more of unit performance diagnosis and prediction algorithm management.

[0029] In addition, the above DT computation engine can receive computational assistance through one or more of the analysis algorithm manager, prediction algorithm manager, prediction engine, analysis engine, torrefaction data transmission and reception unit and H&MB data transmission and reception unit.

[0030] In addition, the above AI modeling engine can continuously improve the model by additionally including verification means.

[0031] In addition, any one or more of the above database and DT computing engine can collect additional data using an external platform.

[0032] In addition, the above external platform may be one or more of Kafka and Zookeeper.

[0033] In addition, the above plant system can further utilize normalized vision data.

[0034] In addition, the digital twin-based renewable fuel plant system provides a method for operating a digital twin-based renewable fuel plant system comprising: a first step of measuring first data; a second step of storing the measured first data in a large-scale high-speed database; a third step of generating second data through an AI modeling engine based on the first data stored through the second step; a 4-1 step of providing the generated second data to an intelligent central control system; a 4-2 step of providing unit performance diagnosis information to the intelligent central control system; a 4-3 step of providing prediction algorithm management information to the intelligent central control system; a fifth step in which the intelligent central control system performs 3D visualization through a 3D visualization engine based on the received data; and a computation step of performing computations on one or more of the database, the intelligent central control system, and the 3D visualization engine. Effects of the invention

[0036] The present invention can provide a system that generates virtual data for modeling and simulation by selecting a prediction algorithm. Brief explanation of the drawing

[0038] Figure 1 shows the configuration of a digital twin-based renewable fuel plant system, which is an embodiment of the present invention. Figure 2 shows information for performing a data statistical prediction model as an embodiment of the present invention. FIG. 3 is a flowchart for selecting an optimal algorithm as an example of the present invention. Specific details for implementing the invention

[0039] The present invention will be described in more detail below with reference to the attached drawings. However, the attached drawings are merely examples to easily explain the content and scope of the technical concept of the present invention, and the technical scope of the present invention is not limited or altered by these examples. Furthermore, it will be obvious to those skilled in the art that various modifications and changes are possible within the scope of the technical concept of the present invention based on these examples.

[0040] Furthermore, terms and words used in this specification and claims should not be interpreted as being limited to their ordinary or dictionary meanings, but should be interpreted in a meaning and concept consistent with the technical spirit of the invention, based on the principle that the inventor can appropriately define the concept of the terms to best describe his invention. Accordingly, the embodiments described in this specification are merely the most preferred embodiments of the invention and do not represent all of the technical spirit of the invention; therefore, it should be understood that various equivalents and modifications that can replace them may exist at the time of filing this application.

[0041] To illustrate an embodiment of the present invention, the drawings are as follows. FIG. 1 shows the configuration of a digital twin-based renewable fuel plant system, which is an embodiment of the present invention; FIG. 2 shows information for performing a data statistical prediction model, which is an embodiment of the present invention; and FIG. 3 is a flowchart for selecting an optimal algorithm, which is an embodiment of the present invention.

[0042] The present invention provides a digital twin-based renewable fuel plant system comprising: first data which is measured data; a database that stores the first data in large capacity at high speed; an AI modeling engine that performs an algorithm based on the first data stored in the database to provide second data; an intelligent central control system that receives the second data modeled by the AI ​​modeling engine; a 3D visualization engine that visualizes the received data in 3D in the intelligent central control system; and a DT computation engine that performs computations on one or more of the database, the intelligent central control system, and the 3D visualization engine.

[0043] In addition, the first data may be any one or more of the following: primary combustion furnace TC, secondary combustion furnace TC, combustion furnace outlet temperature, B / F temperature, differential pressure, pulsing history, pellet input amount, calorific value, ternary, elemental adjustment, combustion furnace temperature, NOx, SOx, O2, and combustion chamber flame image.

[0044] In addition, the above AI modeling engine may include a data statistical prediction model.

[0045] The above data statistical prediction model clusters the received signals so that the AI ​​model can learn on a unit basis.

[0046] In addition, the data statistical prediction model can calculate the degree of similarity after the AI ​​model compares the learned unit with the current state and waveform on a unit-by-unit basis.

[0047] In addition, the above intelligent central control system can perform one or more of unit performance diagnosis and prediction algorithm management.

[0048] In addition, the above prediction algorithm management may select one or more of the multiple prediction algorithms.

[0049] In addition, the above prediction algorithm management can be updated.

[0050] In addition, the above DT computation engine can receive computational assistance through one or more of the analysis algorithm manager, prediction algorithm manager, prediction engine, analysis engine, torrefaction data transmission and reception unit and H&MB data transmission and reception unit.

[0051] The above H&MB stands for Heat & Material Balance.

[0052] In addition, the above AI modeling engine can continuously improve the model by additionally including verification means.

[0053] In addition, the above prediction / analysis engine can perform one or more of H&MB, carbonization degree prediction, carbonization degree prediction, moisture content prediction, and calorific value prediction.

[0054] In addition, the verification means may utilize information provided by the data statistical prediction model.

[0055] In addition, the verification means may include means for monitoring the waveform in real time and storing an abnormal signal simultaneously with an alarm if it deviates from a set similar level.

[0056] In addition, the verification means may include means for monitoring operation and status values ​​in real time and quantifying and storing the degree of deviation when it deviates from a set similar level.

[0057] In addition, any one or more of the above database and DT computing engine can collect additional data using an external platform.

[0058] In addition, the above external platform may be one or more of Kafka and Zookeeper.

[0059] In addition, the above plant system can further utilize normalized vision data.

[0060] In addition, the above vision data may be data that acquires an image of the used ash.

[0061] The above vision data consists of three types: one for measuring flames, another for measuring chars for feedback control, and the last one for measuring ash. So, after first creating a normalization that adjusts data with a resolution of 255 or 1024 to 100, the normalized RGB values ​​are combined to form a single time series data, and this time series data can enable energy-saving operation through feedback via linear prediction.

[0062] The above normalized vision data may be data including RGB, and the RGB may include all data using the three primary colors, Red, Green, and Blue.

[0063] In addition, the normalized vision data may additionally include means for normalizing data using the three primary colors, or may utilize pre-normalized data.

[0064] In addition, the normalized vision data may be, for example, data in which 256 colors are divided into a range from 1 to 100, and the division from 1 to 100 is merely an example and can be flexibly changed and applied depending on the situation and raw materials.

[0065] In addition, the moisture content of raw materials or products can be measured and predicted through the above-mentioned normalized vision data.

[0066] In addition, a high moisture content is disadvantageous in terms of density because it retains a large amount of water, whereas the less moisture there is, the higher the density, making it suitable for use as a good torrefied fuel.

[0067] In addition, for this process, the speed of the conveyor belt motor that feeds fuel is controlled, the burner of the combustion furnace is controlled through dual PID control, the speed of the screw (feeder) inside the combustion furnace is controlled, and the condition of the ash is measured through a vision system and the operating value after one minute is predicted to save energy and enable optimal control operation.

[0068] Furthermore, torrefied materials that have undergone the torrefaction process have lower moisture content and an O / C (oxygen / carbon) ratio, and can possess high energy density and calorific value suitable for fuel. (Moisture decreases during torrefaction (moisture content decreases)) More precisely, the oxygen and hydrogen components decrease and change into carbon properties. A vision-based prediction engine can be used to control this.

[0069] In addition, the normalized vision data can be used in one or more of the database, AI modeling engine, intelligent central control system, 3D visualization engine, and DT computation engine.

[0070] In addition, it can also be used during the timing of raw material input.

[0071] In addition, the digital twin-based renewable fuel plant system may include the selection of an optimal algorithm through a sequence of selecting a prediction algorithm for verification of the prediction algorithm, generating virtual data for modeling, loading virtual environment data and visualizing it as a 3D graphic, generating execution results of the selected prediction algorithms, comparing the execution results for each algorithm with the equipment data at the site, selecting the optimal algorithm in each time-clustered section based on the comparison results, and displaying the optimal algorithm based on the 3D-based comparison results.

[0072] In addition, the digital twin-based renewable fuel plant system provides a method for operating a digital twin-based renewable fuel plant system comprising: a first step of measuring first data; a second step of storing the measured first data in a large-scale high-speed database; a third step of generating second data through an AI modeling engine based on the first data stored through the second step; a 4-1 step of providing the generated second data to an intelligent central control system; a 4-2 step of providing unit performance diagnosis information to the intelligent central control system; a 4-3 step of providing prediction algorithm management information to the intelligent central control system; a fifth step in which the intelligent central control system performs 3D visualization through a 3D visualization engine based on the received data; and a computation step of performing computations on one or more of the database, the intelligent central control system, and the 3D visualization engine.

[0073] In addition, the above steps 4-1, 4-2, and 4-3 may proceed in order or independently regardless of order.

[0074] In addition, the above operation step can be performed regardless of the order of progress of steps 1 through 5.

[0075] In addition, a data statistical prediction model may be used in the third step mentioned above.

[0076] In addition, the above data statistical prediction model may include a step of clustering the received signals and having the AI ​​model learn on a unit basis.

[0077] In addition, the data statistical prediction model may include a step in which the AI ​​model compares the learned unit with the current state and waveform on a unit-by-unit basis and calculates the degree of similarity.

[0078] In addition, the above data statistical prediction model may include a verification step that verifies through a verification method.

[0079] In addition, the verification step may include a step of monitoring the waveform in real time and saving an abnormal signal simultaneously with an alarm if it deviates from a set similar level.

[0080] In addition, the verification step may include a step of monitoring operation and status values ​​in real time and quantifying and storing the degree of deviation when it deviates from a set similar level.

[0081] In addition, the above system may further include a step of selecting an optimal algorithm based on the sequence for selecting an optimal algorithm.

[0082] In addition, the above-mentioned digital twin-based renewable fuel plant system or system operation method may utilize biomass as fuel.

[0083] In addition, to utilize the above biomass, it may be applied to a gasification device comprising, in one embodiment, a solid generation process that receives biomass and heats it to produce a solid product through combustion, carbonization, and torrefaction; a gasification reaction process that receives the solid product produced in the solid generation process and produces a synthesis gas containing one or more of hydrogen or carbon monoxide; and an air separation process that separates nitrogen from air.

[0084] Although the present invention has been described in detail through representative embodiments above, those skilled in the art will understand that various modifications can be made to the above-described embodiments within the scope of the present invention.

[0085] Therefore, the scope of the present invention should not be limited to the described embodiments, but should be determined by the claims set forth below as well as equivalents thereof.

Claims

Claim 1 A digital twin-based renewable fuel plant system comprising: first data which is measured data; a database that stores the first data in large capacity at high speed; an AI modeling engine that performs an algorithm based on the first data stored in the database to provide second data; an intelligent central control system that receives the second data modeled by the AI ​​modeling engine; a 3D visualization engine that visualizes the received data in 3D in the intelligent central control system; and a DT computation engine that performs operations on one or more of the database, the intelligent central control system, and the 3D visualization engine; wherein one or more of the database and the DT computation engine additionally collect data using an external platform, and the external platform uses one or more of Kafka and Zookeeper. Claim 2 A digital twin-based renewable fuel plant system according to claim 1, wherein the first data is any one or more of a primary combustion furnace TC, a secondary combustion furnace TC, a combustion furnace outlet temperature, a B / F temperature, a differential pressure, a pulsing history, a pellet input amount, a calorific value, a triad, an elemental adjustment, a combustion furnace temperature, NOx, SOx, O2, and a combustion chamber flame image. Claim 3 In claim 1, the AI ​​modeling engine is a digital twin-based renewable fuel plant system including a data statistical prediction model. Claim 4 In paragraph 1, the intelligent central control system is a digital twin-based renewable fuel plant system that performs one or more of unit performance diagnosis and prediction algorithm management. Claim 5 In claim 1, the DT computation engine is a digital twin-based renewable fuel plant system that receives computational assistance through one or more of an analysis algorithm manager, a prediction algorithm manager, a prediction engine, an analysis engine, a torrefaction data transmission and reception unit, and an H&MB data transmission and reception unit. Claim 6 In claim 1, the AI ​​modeling engine further includes a verification means, resulting in a digital twin-based renewable fuel plant system in which the model is continuously improved. Claim 7 A digital twin-based renewable fuel plant system according to claim 1, wherein at least one of the database and the DT computing engine additionally collects data using an external platform. Claim 8 delete Claim 9 In claim 1, the plant system is a digital twin-based renewable fuel plant system that additionally utilizes normalized vision data. Claim 10 A method for operating a digital twin-based renewable fuel plant system according to claim 1, comprising: a first step of measuring first data; a second step of storing the measured first data in a large-scale high-speed database; a third step of generating second data through an AI modeling engine based on the first data stored through the second step; a 4-1 step of providing the generated second data to an intelligent central control system; a 4-2 step of providing unit performance diagnosis information to the intelligent central control system; a 4-3 step of providing prediction algorithm management information to the intelligent central control system; a fifth step in which the intelligent central control system performs 3D visualization through a 3D visualization engine based on the received data; and a computation step of performing computations on one or more of the large-scale high-speed database, the intelligent central control system, and the 3D visualization engine.

Images