Carbon cycle system based on automatic control

Abstract

The present invention discloses a carbon cycle system based on automatic control, which relates to the technical field of carbon cycle; the system includes: a power generation unit, a hydrogen production unit, a methane synthesis unit, a natural gas power unit, a carbon capture unit and an automatic control system; the power generation unit transmits the water generated by power generation to the hydrogen production unit, and the hydrogen production unit electrolyzes water and transmits the generated hydrogen to the methane synthesis unit; the natural gas power unit transmits the generated gas to the storage unit, and the carbon capture unit captures carbon dioxide from the gas in the storage unit and then transmits the carbon dioxide to the methane synthesis unit; the methane synthesis unit generates methane based on hydrogen and carbon dioxide and then transmits the generated methane to the natural gas power unit; the present invention has the advantages of intelligent management, accurate prediction, closed-loop cycle, flexible scalability, etc., and can effectively solve the problems existing in traditional carbon capture technologies.

Description

Technical Field

[0001] The present invention relates to the technical field of carbon cycle, and particularly to a carbon cycle system based on automatic control. Background Art

[0002] With the continuous growth of global energy demand and the increasingly prominent problem of climate change, the demand for clean energy and low-carbon technologies is becoming increasingly urgent. In this context, carbon cycle technology, as an effective means to mitigate climate change, has attracted much attention. Carbon cycle technology aims to capture and utilize greenhouse gases such as carbon dioxide from the atmosphere / tail gas, thereby reducing its impact on the Earth's climate system and realizing the recycling of carbon resources. In carbon cycle technology, carbon capture is a crucial link, which involves the process of capturing carbon dioxide from industrial emission gases or the atmosphere. However, traditional carbon capture technologies usually have problems such as high cost and large energy consumption, which limit their wide application in industrial practice. Therefore, it is necessary to develop new carbon capture technologies to solve the limitations of existing technologies.

[0003] Currently, there are already some carbon capture technologies, such as physical adsorption, chemical absorption, and membrane separation. Physical adsorption technology uses adsorbents to adsorb carbon dioxide to separate gas mixtures, chemical absorption technology separates carbon dioxide by the chemical reaction between the absorbent in the solution and carbon dioxide, and membrane separation technology uses the difference in the diffusion rate of gases on the membrane to achieve gas separation. Although these technologies can achieve carbon capture to a certain extent, there are still some problems:

[0004] Firstly, traditional carbon capture technologies often use adsorbents or absorbents with high energy consumption, which leads to an increase in the operating cost of equipment. For example, in the chemical absorption process, commonly used amine absorbents have disadvantages such as high energy consumption and easy volatilization, making the carbon capture cost relatively high. Secondly, the equipment in traditional carbon capture technologies is complex, large in volume, and poor in operating stability, which is not conducive to large-scale industrial applications. For example, physical adsorption and chemical absorption technologies require the construction of large equipment for adsorbing or absorbing carbon dioxide, while although the equipment of membrane separation technology is relatively small, there are problems such as the selection and lifespan of membrane materials, which affect its long-term stable operation. In addition, traditional carbon capture technologies also have certain limitations in terms of capture efficiency and reuse of carbon dioxide. For example, some traditional technologies have low capture efficiency for carbon dioxide, and the captured carbon dioxide is also difficult to be effectively reused, resulting in waste of carbon resources. Summary of the Invention

[0005] In view of this, the present invention provides a carbon cycle system based on automatic control, which has the advantages of intelligent management, accurate prediction, closed-loop cycle, flexible scalability, etc., and can effectively solve the problems existing in traditional carbon capture technologies, providing new ideas and methods for the development and application of carbon cycle technology.

[0006] The technical solution adopted by the present invention is as follows:

[0007] A carbon cycle system based on automatic control, the system includes: a power generation unit, a hydrogen production unit, a methane synthesis unit, a natural gas power unit, a carbon capture unit and an automatic control system; the power generation unit is used for power generation and is used by the hydrogen production unit; the hydrogen production unit generates hydrogen by electrolyzing water and transmits the generated hydrogen to the methane synthesis unit; the natural gas power unit uses the methane generated by the methane synthesis unit to produce continuous power or electricity and generates carbon dioxide; the carbon capture unit captures carbon dioxide from the tail gas generated by the natural gas power unit and then transmits the carbon dioxide to the methane synthesis unit; the methane synthesis unit generates methane based on hydrogen and carbon dioxide and then transmits the generated methane to the natural gas power unit; the automatic control system includes: an automatic control unit, a sensor group and an electric valve group, the electric valve group includes a plurality of electric valves respectively arranged at the output ports of the power generation unit, the hydrogen production unit, the methane synthesis unit, the natural gas power unit and the carbon capture unit, and each electric valve responds to the control command of the automatic control unit to control the opening and closing of the electric valve so as to maximize the carbon cycle efficiency of the system; the sensor group includes a plurality of sensors respectively arranged inside the hydrogen production unit, the methane synthesis unit, the natural gas power unit and the carbon capture unit, and real-time collects data and sends it to the automatic control unit.

[0008] Further, the automatic control unit includes: a data processing unit, a state prediction unit and a control unit; the data collected by the sensor group at least includes: the hydrogen content in the methane synthesis unit , the carbon dioxide content and the methane production rate , the carbon dioxide content in the carbon capture unit , the water volume in the hydrogen production unit and the hydrogen production rate ; the data processing unit is used for preprocessing the data collected by the sensor group and then constructing a system state vector :

[0009] ;

[0010] wherein, is the data collected by the sensor group, ; input the system state vector into the state prediction unit; after receiving the system state vector, the state prediction unit generates a predicted system state vector at the next moment; the control unit, for the predicted system state vector, finds the preset template vector closest to it, and according to the closest preset template vector, calls the corresponding control instruction to control the opening and closing of each electric valve in the electric valve group.

[0011] Further, the method for generating the predicted system state vector at the next moment after the state prediction unit receives the system state vector includes:

[0012] Step 1: Generate the observed system state vector at each moment according to the system state vector;

[0013] Step 2: Input the observed system state vector into a preset improved unscented Kalman filter to generate the predicted system state vector and obtain the covariance matrix of the predicted system state;

[0014] Step 3: Compare the observed system state vector with the predicted system state vector to update the improved unscented Kalman filter of the system.

[0015] Further, in Step 1, the observed system state vector at each moment is generated according to the system state vector by the following formula:

[0016] ;

[0017] where is the preset observation matrix, is the observation noise, which follows a Gaussian distribution.

[0018] Further, the improved unscented Kalman filter in Step 2 is represented by the following formula:

[0019] ;

[0020] where is the predicted system state vector at time , is the covariance matrix of the predicted system state at time , is the state transition function, is the th weighting coefficient, is the th sigma point, is the covariance matrix of the process noise; is the dimension of the state vector, is the scale parameter; is the transpose operation of the matrix; represents the standard deviation of the th uncertainty; is the predicted system state vector at time ; is the covariance matrix of the predicted system state at time .

[0021] Further, the state transition function It is represented by the following formula:

[0022] ;

[0023] where, is the gain of the unscented Kalman filter at the improved moment ; represents the calculation of the second-order Frobenius norm of the matrix.

[0024] Furthermore, step 3 specifically includes: using the following formula to update the gain of the improved unscented Kalman filter:

[0025] .

[0026] where, is the gain of the unscented Kalman filter at the improved moment ; represents the calculation of the determinant of the matrix.

[0027] Furthermore, each preset template vector corresponds to a control instruction; each control instruction is a vector, and each vector element is a binary array; in the binary array, the first number represents the valve opening / closing state, and the second number represents the time of valve opening / closing.

[0028] Furthermore, the method for preprocessing the data collected by the sensor group by the data processing unit includes: performing data fuzzification processing on the data collected by the sensor group, specifically including: generating fuzzy data within a set error range from the collected data through a generative adversarial network.

[0029] With the above technical solutions, the present invention has the following beneficial effects: The present invention uses automatic control technology to intelligently manage and optimize the control of the carbon cycle system, enabling real-time monitoring and adjustment of the system operating status. The operating data of each component is collected in real time by the sensor group in the automatic control system and sent to the automatic control unit for processing and analysis. The automatic control unit calls the corresponding control instructions according to the preset control strategy to control the working status of each component, so as to achieve the optimal operation of the system. This intelligent management and optimization control can improve the operating efficiency and stability of the carbon cycle system, reduce human intervention, and lower the operating cost. Secondly, the present invention adopts advanced algorithms such as the improved unscented Kalman filter to achieve accurate estimation and prediction of the system state. In the carbon cycle system, the state prediction unit uses the improved unscented Kalman filter to predict the system state and generate the predicted system state vector at the next moment. This accurate state prediction can help the system detect potential problems in a timely manner, predict future operating trends, provide accurate decision-making basis for the automatic control system, and improve the operating reliability and safety of the system. Thirdly, the present invention adopts a variety of advanced technologies, such as water electrolysis for gas production, methane synthesis, carbon capture, etc., to achieve a closed-loop cycle of the carbon cycle process. In the present invention, the water generated by the power generation unit is used to produce hydrogen through the water electrolysis for gas production unit, and then reacts with the carbon dioxide captured by the carbon capture unit to synthesize methane and transmit it to the natural gas power unit to form natural gas supply. At the same time, the carbon capture unit captures carbon dioxide from the gas in the storage unit, realizing the recycling of carbon dioxide. This closed-loop carbon cycle process can not only effectively reduce carbon dioxide emissions, but also reduce the dependence on traditional energy sources and achieve the efficient utilization of carbon resources. Finally, the carbon cycle system based on automatic control proposed by the present invention has high flexibility and scalability. In this system, the control instructions can be flexibly adjusted according to the actual situation to adapt to different working conditions and operating requirements. At the same time, the various components in the system adopt a modular design and can be individually upgraded or replaced according to needs, thus realizing the scalability and maintainability of the system. This flexibility and scalability enable the carbon cycle system to adapt to the application requirements of different scales and different fields and have high application potential and market competitiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 FIG. is a schematic flow chart of a mobile phone signal enhancement method based on multi-channel adaptive beamforming in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0031] All features disclosed in this specification, or all steps in any method or process disclosed, can be combined in any manner, except for mutually exclusive features and / or steps.

[0032] Example 1: Refer toFigure 1 A carbon cycle system based on automatic control, the system comprising: a power generation unit, a hydrogen production unit, a methane synthesis unit, a natural gas power unit, a carbon capture unit and an automatic control system; the power generation unit mainly generates green electricity through solar energy, wind energy, etc. for the hydrogen production unit to use. The hydrogen production unit mainly generates green electricity through solar energy, wind energy, etc. for the hydrogen production unit to use; the hydrogen production unit generates hydrogen by means of electrolyzing water, etc., and transmits the generated hydrogen to the methane synthesis unit; the natural gas power unit uses the methane produced by the methane synthesis unit to produce continuous power or electricity, and generates carbon dioxide; the carbon capture unit captures carbon dioxide from the tail gas of the natural gas power unit, and then transmits the carbon dioxide to the methane synthesis unit; the methane synthesis unit generates methane based on hydrogen and carbon dioxide, and then transmits the generated methane to the natural gas power unit; the automatic control system comprises: an automatic control unit, a sensor group and an electric valve group, the electric valve group comprising a plurality of electric valves respectively arranged at the output ports of the power generation unit, the hydrogen production unit, the methane synthesis unit, the natural gas power unit and the carbon capture unit, each electric valve responding to the control command of the automatic control unit to control the opening and closing of the electric valve so as to maximize the carbon cycle efficiency of the system; the sensor group comprises a plurality of sensors respectively arranged inside the hydrogen production unit, the methane synthesis unit, the natural gas power unit and the carbon capture unit, and real-time collects data and sends it to the automatic control unit.

[0033] Specifically, this carbon cycle system based on automatic control is a system that comprehensively utilizes water, hydrogen and carbon dioxide, aiming to achieve high-efficiency carbon cycling and resource utilization. The power generation unit is the starting point of the system, and it generates electrical energy. The power generation unit here can be any kind of power generation equipment, but mainly renewable energy (such as solar energy, wind energy, etc.) power generation devices, so as to maximize the system efficiency. The hydrogen production unit can decompose water into hydrogen and oxygen by using the electrolysis principle. Water is decomposed into hydrogen and oxygen during the electrolysis process, where the principle of electrolysis is to utilize the oxidation-reduction reaction that occurs when an electric current passes through an aqueous solution. The specific reaction equation is: ; The hydrogen generated in this step is used for subsequent methane synthesis. The methane synthesis unit uses the syngas method (usually the Sabatier synthesis) or the biological method to react hydrogen with carbon dioxide to generate methane . The syngas method is an important industrial chemical reaction that uses a certain proportion of hydrogen and carbon dioxide or carbon monoxide as raw materials, and generates methane through a catalytic reaction. The reaction usually takes place under high temperature and high pressure. The chemical formula of the synthesis reaction is as follows: ; The methane produced here can be used as a substitute for natural gas or for other industrial purposes. The natural gas power unit receives and stores the produced gas, the main component of which is methane. Methane is a common clean fuel gas and can be directly used for heating, power generation and other purposes. The storage unit is used to temporarily store the gas produced by the power generation unit and the natural gas power unit to balance the production and consumption between different units in the system. The carbon capture unit is a key part of the system and is used to capture carbon dioxide from the gas produced by the natural gas power unit. This can be achieved by methods such as adsorbents or chemical absorbents. The captured carbon dioxide is then transported to the methane synthesis unit for methane synthesis. The automatic control system includes an automatic control unit, a sensor group and an electric valve group. The sensor group is responsible for collecting data inside the hydrogen production unit, methane synthesis unit, natural gas power unit and carbon capture unit in real time and sending the data to the automatic control unit. The automatic control unit adjusts the operating states of various parts of the system by controlling the electric valves in the electric valve group according to the data collected by the sensors to ensure the efficient operation of the carbon cycle system. The creativity of the whole system lies in its comprehensive utilization of water, hydrogen and carbon dioxide resources, realizing the efficient conversion and utilization of these resources through automatic control, while reducing greenhouse gas emissions, and having high environmental friendliness and economy. Compared with the traditional energy production system with high carbon emissions, this system has significant innovations and advantages in resource utilization and environmental protection. The methane synthesis unit can also be replaced with a methanol synthesis unit or other carbon-hydrogen fuel synthesis units that are convenient for storage and transportation; the natural gas power unit can also be replaced with a methanol or carbon-hydrogen fuel power unit.

[0034] Embodiment 2: The automatic control unit includes: a data processing unit, a state prediction unit and a control unit; the data collected by the sensor group at least includes: the hydrogen content in the methane synthesis unit , the carbon dioxide content and the methane production rate , the carbon dioxide content in the carbon capture unit , the water volume in the hydrogen production unit and the hydrogen production rate ; The data processing unit is used to preprocess the data collected by the sensor group and then construct a system state vector :

[0035] ;

[0036] Wherein, is the data collected by the sensor group, ; The system state vector Input into the state prediction unit; after receiving the system state vector, the state prediction unit generates a predicted system state vector for the next moment; the control unit finds the preset template vector closest to the predicted system state vector and, according to the closest preset template vector, calls the corresponding control instruction to control the opening and closing of each electric valve in the electric valve group.

[0037] Specifically, the data processing unit is responsible for preprocessing the data collected by the sensor group. The preprocessing includes operations such as data cleaning, denoising, and normalization to ensure the quality and usability of the data. After preprocessing, the data is constructed into a system state vector , which includes various data collected by the sensors. The construction of the system state vector enables the state information of the system to be transmitted and processed in a unified form.

[0038] State prediction unit: The state prediction unit receives the system state vector constructed by the data processing unit . It uses prediction algorithms (such as methods based on time series analysis, machine learning, or deep learning) to predict the state of the system at the next moment. This prediction can help the system make adjustments in advance to cope with possible changes and fluctuations, thereby improving the stability and efficiency of the system. The control unit is the decision-making center of the entire automatic control system. It receives the predicted system state vector generated by the state prediction unit and compares it with the preset template vector. The preset template vector is defined in advance and represents the expected performance of the system in different operating states. The control unit selects the preset template vector closest to the predicted system state vector according to the comparison result and calls the corresponding control instruction. The control instruction is usually for the opening and closing control of the electric valve group to adjust the operating states of each unit in the system. The opening and closing control of the electric valve can adjust the flow rate and output between each unit to achieve the dynamic balance and optimization of the system operation. The entire automatic control unit uses advanced data processing and prediction algorithms to achieve real-time monitoring and regulation of the operating state of the carbon cycle system. By predicting the future state and controlling according to the preset template, the system can be made more intelligent and adaptive, thereby improving the stability, efficiency, and reliability of the carbon cycle system. Compared with the traditional manual regulation method, this automatic control unit has higher precision and response speed and can better adapt to the complex and changeable industrial production environment.

[0039] Embodiment 3: The method for the state prediction unit to generate a predicted system state vector for the next moment after receiving the system state vector includes:

[0040] Step 1: Generate an observed system state vector for each moment according to the system state vector;

[0041] The system state vector is a vector that contains various state information of the system at a specific moment. In the carbon cycle system, this state information may include the hydrogen content, carbon dioxide content, and methane production rate in the methane synthesis unit, the carbon dioxide content in the carbon capture unit, the water volume and hydrogen production rate in the hydrogen production unit, etc. Therefore, the system state vector can comprehensively reflect the current operating state of the system. Secondly, the generation of the observed system state vector is completed through the data collected by sensors. A sensor is a device that can sense and measure environmental information. It can collect data from various parts of the system in real time, such as gas concentration, liquid flow rate, etc. These data form the observed system state vector after sampling and processing. It should be noted that the data collected by sensors may be affected by noise and interference. Therefore, data cleaning and processing are required when generating the observed vector to ensure the accuracy and reliability of the data. The process of generating the observed system state vector is a real-time and continuous process. As time goes by, both the system state and sensor data are constantly changing. Therefore, the observed system state vector will also be updated accordingly. This enables the state prediction unit to perform state prediction based on the latest observed data, thus more accurately reflecting the dynamic changes of the system. Generally speaking, step 1 is to use the data collected by sensors, combined with the system state vector, to generate an observed vector that reflects the current state of the system. This process provides key input data for the state prediction unit and lays the foundation for subsequent state prediction and control. By continuously updating the observed vector, the system can achieve real-time monitoring and feedback of the current state, thereby realizing the effective control and management of the carbon cycle system.

[0042] Step 2: Input the observed system state vector into a pre-set improved unscented Kalman filter to generate a predicted system state vector and obtain the covariance matrix of the predicted system state;

[0043] The improved Unscented Kalman Filter is a variant of the Kalman Filter, which is mainly used to estimate the state of a nonlinear system. Compared with the standard Kalman Filter, the improved Unscented Kalman Filter transforms the uncertainty of the nonlinear system into a Gaussian distribution through the unscented transform (such as the Sigma-Point transform), thus better handling the problems of nonlinear systems and uncertainties. Next, understand the working process of the filter. In Step 2, the observed system state vector is input into the improved Unscented Kalman Filter. The filter uses the observed data at the current moment, combines the system state transition equation and the dynamic model, and generates the predicted system state vector at the next moment through recursive calculation. At the same time, the filter also calculates the covariance matrix of the predicted system state, which is used to measure the degree of uncertainty of the predicted state. In the improved Unscented Kalman Filter, the main operations include: Prediction Step (PredictStep): According to the system state transition equation and the dynamic model, use the current state to predict the state of the system at the next moment. This step usually uses numerical integration methods (such as the Euler method or the Runge-Kutta method) for numerical calculation. Update Step (UpdateStep): Use the observed data to correct the predicted state and update the predicted state estimate. In this step, the filter takes into account the uncertainty of the observed data and fuses the observed data with the predicted data to improve the accuracy and precision of the state estimate. In Step 2, the improved Unscented Kalman Filter outputs the predicted system state vector and the covariance matrix of the predicted system state. The predicted system state vector contains the predicted state information of the system at the next moment, while the covariance matrix provides the uncertainty information of the predicted state. Step 2 is to predict the future state of the system using the improved Unscented Kalman Filter. Through recursive calculation and state correction, the filter can effectively handle nonlinear systems and uncertainties, thereby improving the prediction accuracy and precision of the system state.

[0044] Step 3: Compare the observed system state vector with the predicted system state vector and update the improved Unscented Kalman Filter of the system.

[0045] The observation system state vector is generated based on the actual data collected by sensors, reflecting the true state of the system at the current moment. The predicted system state vector is obtained by predicting with an improved unscented Kalman filter, representing the predicted state of the system at the next moment. By comparing these two vectors, the accuracy of the prediction can be evaluated, and the parameters of the filter can be adjusted accordingly to improve the accuracy and reliability of the state prediction. Secondly, the method of comparing the observation system state vector and the predicted system state vector usually involves calculating the difference or distance between them. Commonly used comparison methods include Euclidean distance, Mahalanobis distance, correlation coefficient, etc. Through these methods, the similarity or difference degree between the observed data and the predicted data can be quantified, so as to determine whether it is necessary to adjust the parameters of the filter or update the state estimation. In actual operation, the comparison result is usually compared with a pre-set threshold. If the comparison result exceeds the threshold, it means that there is a large difference between the predicted system state and the observed system state, and the filter needs to be adjusted or updated. Otherwise, it can be considered that the predicted system state is basically consistent with the observed system state, and no additional processing is required. Finally, the improved unscented Kalman filter is updated according to the comparison result. According to the comparison result, the parameters of the filter can be adjusted or the state estimation can be updated to make the predicted system state closer to the observed system state. This step usually involves updating the state estimation and covariance matrix of the filter to reflect the latest observed data and prediction results. To sum up, step 3 is to evaluate the accuracy of the prediction by comparing the observation system state vector and the predicted system state vector, and update the improved unscented Kalman filter according to the comparison result. This process can help the system dynamically adjust parameters to achieve more accurate prediction and more effective control of the system state, thereby improving the operation efficiency and stability of the carbon cycle system.

[0046] Example 4: Step 1, according to the following formula, generate the observation system state vector at each moment based on the system state vector:

[0047] ;

[0048] where is the preset observation matrix, is the observation noise, which follows a Gaussian distribution.

[0049] Specifically, in Example 4, step 1 involves using the formula to generate the observation system state vector at each moment. The system state vector is a vector containing various state information of the system at a specific moment. In the carbon cycle system, it may include the hydrogen content, carbon dioxide content, and methane production rate in the methane synthesis unit, the carbon dioxide content in the carbon capture unit, the water volume and hydrogen production rate in the hydrogen production unit, etc. The observation matrix is a pre - set matrix used to transform the system state vector into the observed system state vector. The selection of is determined according to the characteristics and requirements of the specific system. It is usually a linear transformation matrix used to extract part of the information in the system state vector to construct the observed system state vector. Observation noise refers to the random error or perturbation in the observed system state vector, which usually follows a Gaussian distribution. The existence of this noise is due to the possible influence of environmental interference or measurement error during sensor data acquisition, so it needs to be considered in the model. The formula represents the observed system state vector which is obtained by linearly transforming the system state vector through the observation matrix , and adding the observation noise . This process can be understood as using the sensor to measure part of the information of the system state vector and considering the random error during the measurement process. According to the formula, we can obtain the observed system state vector at each moment by multiplying the system state vector by the observation matrix and adding the observation noise . This process can transform the information in the system state vector into the observed system state vector and consider the uncertainty in the observation process. The application of this formula enables us to obtain the observed system state vector based on the system state vector, providing a basis for subsequent state prediction and control. Through the generation of the observed system state vector, we can monitor the state of the system in real - time and use this information for state prediction and control in subsequent steps to achieve intelligent management and optimal operation of the carbon cycle system. The principle of the formula is to transform the system state vector into the observed system state vector through the observation matrix

[0050] Example 5: The improved unscented Kalman filter in step 2 is represented by the following formula:

[0051] ;

[0052] where is the predicted system state vector at time , is the covariance matrix of the predicted system state at time , is the state transition function, is the th weighting coefficient, is the sigma points, is the covariance matrix of the process noise; is the dimension of the state vector, is the scaling parameter; is the transpose operation of the matrix; denotes the standard deviation of the th uncertainty; is the predicted system state vector at time is the time covariance matrix of the predicted system state.

[0053] Specifically, in the formula, denotes the predicted system state vector at time which is an estimated value obtained by predicting the predicted system state vector at time through the state transition function and the system state vector at the current time . This state transition function describes how the system state evolves over time and is determined based on the dynamic characteristics and operating rules of the system. In the formula, denotes the covariance matrix of the predicted system state at time which is used to measure the uncertainty degree of the predicted system state. It is obtained by sampling and weighting the predicted system state vector. Specifically, by the weighting coefficient and a set of sigma points sample the predicted system state vector , and then calculate the covariance matrix according to the sampling results. This process takes into account the uncertainty of the predicted system state, making the prediction results more reliable and accurate. In addition, in the formula, the process noise covariance matrix is also introduced to describe the random perturbation or noise in the system state transition process. The process noise covariance matrix is usually a positive definite symmetric matrix, reflecting the uncertainty of the system state change. By adding the process noise covariance matrix to the covariance matrix of the predicted system state, the formula takes into account the uncertainty and change of the system state. The scaling parameter and the state vector dimension in the formula are used to adjust the distribution of the sigma points. The choice of the scaling parameter affects the weight calculation of the sigma points and the accuracy of the sampling results. The state vector dimension determines the number of sigma points and the complexity of the sampling, and is usually associated with the state dimension of the system. In summary, the formula describes the improved unscented Kalman filter at time The prediction process includes predicting the system state vector the calculation of and the covariance matrix of the predicted system state the estimation. By considering factors such as the state transition function, weighted sampling, process noise, and scale parameters, the formula can accurately predict the system state at the next moment and estimate the uncertainty of the state. This prediction process provides an important basis for subsequent state updates and control, enabling the carbon cycle system to achieve intelligent operation and optimal control.

[0054] Example 6: State transition function It is represented by the following formula:

[0055] ;

[0056] where is the gain of the unscented Kalman filter at the improved time ; represents the second-order Frobenius norm of the matrix.

[0057] Specifically represents the state transition function at time , which is used to describe how the system state vector evolves over time. It receives the predicted system state vector at the previous moment and the system state vector at the current moment as inputs and outputs the predicted system state vector at the current moment.

[0058] In the formula is the gain of the unscented Kalman filter at the improved time , which is used to adjust the influence degree of the predicted system state vector at the previous moment during the state transition process. The selection of the gain is usually determined according to the dynamic characteristics of the system and the measurement error, and its role is to enable the state transition function to better correct the predicted system state vector to adapt to the actual changes of the system. In the formula represents the square of the second-order Frobenius norm of the matrix . The Frobenius norm of the matrix is the square root of the sum of the squares of the matrix elements, which represents the size of the matrix. By calculating the square of the second-order Frobenius norm of the matrix , a constant can be obtained, which is used to adjust the influence degree of the system state vector at the current moment during the state transition process. This constant can be regarded as a scaling factor, which is used to balance the contributions of the predicted system state vector at the previous moment and the system state vector at the current moment during the state transition process. Formula Describes the state transition function in the improved unscented Kalman filter for predicting the evolution process of the system state vector at time By adjusting the gain and the square of the second-order Frobenius norm of the matrix , the formula can flexibly calculate the predicted system state vector at the current time based on the predicted system state vector at the previous time and the system state vector at the current time. This process provides an important mathematical tool for state prediction, enabling the carbon cycle system to more accurately predict future states and perform corresponding control and adjustment.

[0059] Example 7: Step 3 specifically includes: Using the following formula to update the gain of the improved unscented Kalman filter:

[0060] .

[0061] Where is the gain of the improved unscented Kalman filter at time ; represents the determinant operation of the matrix.

[0062] Specifically, in Example 7, the formula describes the update process of the gain in the improved unscented Kalman filter. This update process adjusts the gain based on the covariance matrix of the predicted system state at the current time, the observation matrix, and the observation noise covariance matrix to make the estimation of the system state more accurate and reliable. In the formula, represents the gain at time , which is used to correct the difference between the predicted system state vector and the observed system state vector, thereby achieving a more accurate estimation of the system state. The calculation of the gain is based on the determinant operation of the matrix. Specifically, it is the determinant operation on a matrix composed of the covariance matrix of the predicted system state, the observation matrix, and the observation noise covariance matrix to obtain a scalar value. The magnitude of this scalar value reflects the degree of correlation between the predicted system state and the observed system state. If the correlation between the predicted system state and the observed system state is strong, then this scalar value will be relatively large; conversely, if their correlation is weak, then this value will be relatively small. Therefore, this scalar value obtained through the determinant operation can be used to adjust the magnitude of the gain so that it can better adapt to the current observation situation and the uncertainty of the predicted system state. In the formula, represents the gain at time , which is the gain value obtained based on the observation results at the previous time. By multiplying the scalar value obtained through the determinant operation by the gain value at the previous time, the updated gain 。This process enables the gain to be dynamically adjusted according to the observed data at the current moment and the gain value at the previous moment, so that the estimation of the system state can be more accurate and robust. The formula describes the update process of the gain in the improved unscented Kalman filter, which is based on the covariance matrix of the predicted system state, the observation matrix, and the observation noise covariance matrix at the current moment. By performing a determinant operation on the matrix, a scalar value reflecting the correlation of the system state is obtained, and then this scalar value is multiplied by the gain value at the previous moment to obtain the updated gain value. This process enables the gain to be adaptively adjusted according to the actual observed data and the uncertainty of the predicted system state, thereby improving the estimation accuracy and robustness of the system state.

[0063] Example 8: Each preset template vector corresponds to a control instruction; each control instruction is a vector, and each vector element is a binary array; in the binary array, the first number represents the valve opening and closing state, and the second number represents the time of valve opening and closing.

[0064] Specifically, assume that the carbon capture unit is a key component in the carbon cycle system, and its main function is to capture carbon dioxide from natural gas. To achieve refined control of the carbon capture unit, the system predefines several possible working states and designs corresponding control instructions for each state.

[0065] Normal capture state: In this state, the carbon capture unit should capture carbon dioxide with maximum efficiency. The system presets a template vector that includes the desired capture effect. To achieve this effect, the system sets a set of control instructions: Control instruction 1: The valve opening and closing state is open, and the opening and closing time is 60 minutes to ensure continuous carbon dioxide capture. Control instruction 2: The valve opening and closing state is open, and the opening and closing time is 30 minutes to ensure periodic removal of the accumulation in the capture unit.

[0066] Low-load capture state: In this state, the carbon capture unit captures carbon dioxide with lower efficiency to save energy or reduce costs. The system presets another template vector that includes the adaptively adjusted capture effect. To achieve this effect, the system sets a new set of control instructions: Control instruction 1: The valve opening and closing state is open, and the opening and closing time is 45 minutes to reduce energy consumption. Control instruction 2: The valve opening and closing state is open, and the opening and closing time is 15 minutes to ensure periodic removal of the accumulation in the capture unit.

[0067] The purpose of this design is to achieve fine control over each key component in the carbon cycle system. By presetting template vectors, the system can define in advance the expected performance under various working conditions and set corresponding control instructions for each state. These control instructions are represented in the form of vectors, where each vector element corresponds to a control action, and each control action consists of a binary array containing the opening and closing states of the valves and the opening and closing times. For example, assume there are multiple key components in the carbon cycle system, such as a power generation unit, a hydrogen production unit, a methane synthesis unit, etc. For each component, the system can define a series of possible working conditions, such as normal operation, stop operation, low-load operation, etc. For each state, the system can preset a template vector containing the corresponding expected performance. Then, for each template vector, the system can set a set of control instructions to adjust the working states of the various components of the system to achieve the desired operating effect. For example, for the hydrogen production unit, the system can define the following working conditions: normal electrolysis, stop electrolysis, low electrolysis rate, etc. For each state, the system can preset a template vector containing the expected electrolysis effect. Then, for each template vector, the system can set a set of control instructions, such as controlling parameters such as the voltage, current, and electrolyte flow rate of the electrolytic cell, to achieve the expected electrolysis effect.

[0068] Example 9: The method for the data processing unit to preprocess the data collected by the sensor group includes: performing data fuzzification processing on the data collected by the sensor group, specifically including: generating fuzzy data within a set error range from the collected data through a generative adversarial network.

[0069] Specifically, the process of generating fuzzy data within a set error range from the collected data through a generative adversarial network specifically includes:

[0070] Step 1: Initialize the generator and discriminator networks. Initialize the weights of the generator network and the discriminator network as Gaussian noise with a random distribution.

[0071] Step 2: Train the generative adversarial network. For each iteration , randomly select a batch of real data samples from the dataset . The generator network uses the input noise vector to generate fake data samples similar to the real data . The discriminator network evaluates the real data samples and the fake data samples respectively and calculates the loss function , including the cross-entropy loss and the gradient penalty term:

[0072] 。

[0073] Step 3: Update the discriminator network parameters. Calculate the gradients of the discriminator network parameters 。Update the discriminator network parameters using the gradient descent method : 。

[0074] Step 4: Generator network training. The generator network generates new fake data samples. Calculate the loss function of the generator network ,which includes the similarity loss of the generated data and the gradient penalty term:

[0075] 。

[0076] Step 5: Update the generator network parameters. Calculate the gradients of the generator network parameters 。Update the generator network parameters using the gradient descent method : 。

[0077] Step 6: Add data blurring operation. Perform advanced blurring on the fake data samples generated by the generator network ,such as Gaussian mixture blurring or non-linear blurring.

[0078] Step 7: Retrain the discriminator network. Retrain the discriminator network using the blurred fake data and real data 。Repeat steps 2 and 3 to update the discriminator network parameters 。

[0079] Step 8: Retrain the generator network. Use the retrained discriminator network to evaluate the fake data generated by the generator network 。Calculate the loss function of the generator network 。Repeat steps 4 and 5 to update the generator network parameters 。

[0080] Step 9: Iterative training. Repeat steps 6 to 8 until the predetermined number of iterations is reached or the loss function converges.

[0081] Step 10: Generate blurred data. Use the trained generator network to generate blurred data. Output the generated blurred data for subsequent use.

[0082] First, data fuzzification makes data more diverse and generalizable by introducing a certain degree of noise and variation. The real data collected by sensors is often affected by environmental and device factors, with a certain degree of uncertainty and randomness. Through fuzzification, data samples can be made more diverse, covering a wider data distribution, thereby improving the system's adaptability to uncertainty. Second, data fuzzification helps reduce the impact of data noise and interference on system performance. During the process of sensor data collection, it is often interfered by various noises and interferences, such as electromagnetic interference, device vibration, etc. These interferences will affect the accuracy and reliability of the data and reduce the stability of the system. Through fuzzification, the noise and interference in the data can be smoothed and suppressed to a certain extent, reducing the impact on the system and improving the system's robustness. In addition, data fuzzification can also improve the security and privacy of the system. In the carbon cycle system, some key data may involve trade secrets or personal privacy and need to be protected. Through fuzzification, the data does not contain specific detailed information, thereby reducing the risk of real data leakage and improving the security and privacy of the system.

[0083] The present invention is not limited to the foregoing specific embodiments. The present invention extends to any new feature disclosed in this specification or any new combination thereof, as well as any new method or process step disclosed or any new combination thereof.

Claims

1. A carbon cycle system based on automatic control, characterized in that: The system comprises: a power generation unit, a hydrogen production unit, a methane synthesis unit, a natural gas power unit, a storage unit, a carbon capture unit and an automatic control system; the power generation unit is used to generate electricity for use by the hydrogen production unit; the hydrogen production unit generates hydrogen by electrolyzing water, and transmits the generated hydrogen to the methane synthesis unit; the natural gas power unit uses the methane generated by the methane synthesis unit to produce continuous power or electricity, and generates carbon dioxide; the carbon capture unit captures carbon dioxide from the tail gas generated by the natural gas power unit, and then transmits the carbon dioxide to the methane synthesis unit; the methane synthesis unit generates methane based on hydrogen and carbon dioxide, and then transmits the generated methane to the natural gas power unit; the automatic control The control system includes: an automatic control unit, a sensor group and an electric valve group, wherein the electric valve group includes a plurality of electric valves respectively arranged at the output ports of the power generation unit, the hydrogen production unit, the methane synthesis unit, the natural gas power unit and the carbon capture unit, each electric valve responds to the control instruction of the automatic control unit to control the opening and closing of the electric valve so as to maximize the carbon cycle efficiency of the system; the sensor group includes a plurality of sensors respectively arranged inside the hydrogen production unit, the methane synthesis unit, the natural gas power unit and the carbon capture unit, and collects data in real time and sends it to the automatic control unit; after receiving the system state vector, the state prediction unit generates a predicted system state vector at the next moment, comprising: Step 1: Generate the observed system state vector at each moment according to the system state vector; Step 2: Input the observed system state vector into a preset improved unscented Kalman filter to generate a predicted system state vector and obtain a covariance matrix of the predicted system state; Step 3: Compare the observed system state vector with the predicted system state vector and update the system improved unscented Kalman filter; The automatic control unit includes: a data processing unit, a state prediction unit and a control unit; the data collected by the sensor group includes at least: the hydrogen content in the methane synthesis unit , Carbon dioxide content and methane production rate , CO2 content in the carbon capture unit , the amount of water in the hydrogen production unit and hydrogen production rate The data processing unit is used to pre-process the data collected by the sensor group and construct a system state vector : ； in, The data collected by the sensor group, ; The system state vector The state prediction unit generates a predicted system state vector at the next moment after receiving the system state vector; the control unit finds the closest preset template vector to the predicted system state vector, and calls the corresponding control instruction according to the closest preset template vector to control the opening and closing of each electric valve of the electric valve group; Step 1: Generate the observed system state vector at each moment based on the system state vector using the following formula: ； in, is the preset observation matrix, is the observation noise, which follows a Gaussian distribution; Each preset template vector corresponds to a control instruction; each control instruction is a vector, wherein each vector element is a binary array; in the binary array, the first number indicates the valve opening and closing state, and the second number indicates the valve opening and closing time.

2. The carbon cycle system based on automatic control according to claim 1, characterized in that: The improved unscented Kalman filter in step 2 is expressed using the following formula: ； in, It's time The predicted system state vector, It's time The covariance matrix of the predicted system state, is the state transition function, It is The weighting coefficients, It is sigma points, is the covariance matrix of the process noise; is the dimension of the state vector, is the scale parameter; is the transpose operation of the matrix; Indicates The standard deviation of uncertainty; It's time The predicted system state vector of It's time The covariance matrix of the predicted system state.

3. The carbon cycle system based on automatic control according to claim 2, characterized in that: State transfer function Use the following formula to express it: ； in, Time for improvement The gain of the unscented Kalman filter; It means to find the second-order F-norm of the matrix.

4. The carbon cycle system based on automatic control as claimed in claim 3, characterized in that: Step 3 specifically includes: using the following formula to update the gain of the improved unscented Kalman filter: ； in, Time for improvement The gain of the unscented Kalman filter; Represents the operation of finding the determinant of a matrix.

5. The carbon cycle system based on automatic control as claimed in claim 4, characterized in that: The method for preprocessing the data collected by the sensor group by the data processing unit includes: performing data fuzzification processing on the data collected by the sensor group, specifically including: generating fuzzy data within a set error range with the collected data through a generative adversarial network.

Images