Carbon dioxide monitoring method and system for offshore FPSO power station heat pump

Abstract

The present application relates to the technical field of component detection, and discloses a carbon dioxide monitoring method and system for offshore FPSO power station heat pumps. The method comprises: component analysis and heat exchange treatment of FPSO power station flue gas to obtain pre-treated flue gas and pre-treatment parameters, and component detection and water washing purification to obtain first clean flue gas and first treatment data; temperature field analysis and absorption monitoring of the first clean flue gas to obtain second clean flue gas, rich amine liquid and second treatment data; mass transfer detection and deep absorption of the second clean flue gas to obtain third clean flue gas, first lean amine liquid and absorption parameters; energy analysis and regeneration treatment of rich amine liquid to obtain carbon dioxide gas, second lean amine liquid and regeneration parameters; purity analysis and liquefaction treatment of carbon dioxide gas to obtain liquid carbon dioxide and quality parameters. The present application improves the efficiency and accuracy of carbon dioxide monitoring for offshore FPSO power station heat pumps.

Description

Technical Field

[0001] The present application relates to the field of component detection, and in particular to a carbon dioxide monitoring method and system for an offshore FPSO power station heat pump. Background Art

[0002] With the continuous development of offshore oil and gas development technology, floating production storage and offloading (FPSO) has become an important facility for deepwater oil and gas field development. During the operation of FPSO power station, a large amount of The flue gas Not only does it cause a greenhouse effect, it can also corrode equipment. The capture methods include physical absorption, chemical absorption and membrane separation. Among them, the alcohol amine method is widely used because of its simple operation and low cost. The capture process includes three main steps: flue gas pretreatment, absorption and desorption, which are achieved through multi-stage absorption towers and regeneration towers. separation and recycling.

[0003] However, the existing There are many problems when using capture technology at sea. First, due to the limited space on the offshore platform, it is difficult to arrange large tower equipment. Second, the offshore environment is complex and changeable, and the equipment is violently bumpy, resulting in uneven distribution of gas and liquid phases, affecting the absorption efficiency. Third, traditional process monitoring methods mainly rely on single-point sampling and analysis, which cannot achieve Real-time dynamic monitoring of the concentration field results in a lag in process parameter optimization and control. In addition, high energy consumption and high operating costs of the equipment are also important factors restricting its promotion and application. Summary of the invention

[0004] The present application provides a carbon dioxide monitoring method and system for an offshore FPSO power station heat pump. The present application aims to solve the problem that the existing technology is difficult to achieve in an offshore FPSO power station. Real-time dynamic monitoring and intelligent control of the capture process. By establishing a multi-scale and multi-dimensional monitoring network, combined with advanced component detection and intelligent algorithms, Accurate monitoring and optimized control of the entire capture process can improve capture efficiency and reduce energy consumption.

[0005] In a first aspect, the present application provides a carbon dioxide monitoring method for an offshore FPSO power plant heat pump, the carbon dioxide monitoring method for an offshore FPSO power plant heat pump comprising:

[0006] Performing online component analysis on the flue gas of the FPSO power station to obtain flue gas parameter data, performing heat exchange processing on the flue gas and the flue gas parameter data to obtain pre-treated flue gas and pre-treatment parameters;

[0007] Performing component detection processing on the pre-treated flue gas according to the pre-treatment parameters, and performing water washing purification processing on the pre-treated flue gas to obtain first clean flue gas and first processing data;

[0008] Based on the first processing data, the first clean flue gas is subjected to temperature field analysis processing, and the first clean flue gas is subjected to absorption dynamic monitoring processing to obtain second clean flue gas, rich amine liquid, and second processing data;

[0009] Performing mass transfer detection processing on the second clean flue gas according to the second processing data, and performing deep absorption processing on the second clean flue gas to obtain third clean flue gas, first lean amine liquid and absorption parameters;

[0010] Performing energy analysis processing on the rich amine liquid based on the absorption parameters, and performing heat pump regeneration processing on the rich amine liquid to obtain carbon dioxide gas, a second lean amine liquid and regeneration parameters;

[0011] The carbon dioxide gas is subjected to purity analysis according to the regeneration parameters, and the carbon dioxide gas is subjected to compression and liquefaction processing to obtain liquid carbon dioxide and its quality parameters.

[0012] In a second aspect, the present application provides a carbon dioxide monitoring system for an offshore FPSO power plant heat pump, the carbon dioxide monitoring system for an offshore FPSO power plant heat pump comprising:

[0013] An analysis module is used to perform online component analysis on the flue gas of the FPSO power station, obtain flue gas parameter data, perform heat exchange processing on the flue gas and the flue gas parameter data, and obtain pre-processed flue gas and pre-processing parameters;

[0014] a detection module, configured to perform component detection processing on the pre-treated flue gas according to the pre-treatment parameters, and perform water washing and purification processing on the pre-treated flue gas to obtain first clean flue gas and first processing data;

[0015] a processing module, configured to perform temperature field analysis processing on the first clean flue gas based on the first processing data, and perform absorption dynamic monitoring processing on the first clean flue gas to obtain second clean flue gas, rich amine liquid, and second processing data;

[0016] an absorption module, configured to perform mass transfer detection processing on the second clean flue gas according to the second processing data, perform deep absorption processing on the second clean flue gas, and obtain third clean flue gas, first lean amine liquid and absorption parameters;

[0017] a regeneration module, for performing energy analysis processing on the rich amine liquid based on the absorption parameters, performing heat pump regeneration processing on the rich amine liquid, and obtaining carbon dioxide gas, a second lean amine liquid and regeneration parameters;

[0018] The liquefaction module is used to perform purity analysis on the carbon dioxide gas according to the regeneration parameters, and to compress and liquefy the carbon dioxide gas to obtain liquid carbon dioxide and its quality parameters.

[0019] In the technical solution provided by the present application, by performing online component analysis on the flue gas of the FPSO power station and obtaining flue gas parameter data, real-time monitoring of flue gas components is achieved, avoiding the lag of traditional sampling analysis and providing an accurate data basis for subsequent process parameter optimization; by performing heat exchange processing on the flue gas and flue gas parameter data, the system waste heat is fully utilized to improve energy utilization efficiency; the pre-treated flue gas is subjected to component detection and water washing purification treatment according to the pre-treatment parameters to effectively remove impurities in the flue gas and create good conditions for the subsequent absorption process; the first clean flue gas is subjected to temperature field analysis and absorption dynamic monitoring processing based on the first processing data to achieve precise control of the absorption process and improve the absorption efficiency; the second clean flue gas is subjected to mass transfer detection and deep absorption processing according to the second processing data to ensure through multi-stage absorption The system can fully capture the carbon dioxide gas and reduce the escape loss. Based on the absorption parameters, the rich amine liquid is subjected to energy analysis and heat pump regeneration treatment to achieve precise control and optimization of the regeneration energy consumption. According to the regeneration parameters, the carbon dioxide gas is subjected to purity analysis and compression liquefaction treatment to obtain a liquid gas that meets the standards. The whole process has established a complete data collection, processing and control system to achieve Intelligent monitoring and control of the entire capture process significantly improves the stability and reliability of the system operation. At the same time, the combined application of multi-stage absorption and heat pump regeneration effectively reduces the system energy consumption and improves economic efficiency. In addition, the solution fully considers the special environment of the offshore platform, adopts distributed monitoring and adaptive control strategies, can adapt to the influence of adverse factors such as equipment shaking, and ensure the continuous and stable operation of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0021] Figure 1 This is a schematic diagram of an embodiment of a carbon dioxide monitoring method for an offshore FPSO power station heat pump in an embodiment of the present application;

[0022] Figure 2 This is a schematic diagram of an embodiment of a carbon dioxide monitoring system for an offshore FPSO power station heat pump in an embodiment of the present application. DETAILED DESCRIPTION

[0023] The embodiment of the present application provides a carbon dioxide monitoring method and system for an offshore FPSO power plant heat pump. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described here can be implemented in an order other than that illustrated or described here. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0024] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In the embodiment of the present application, an embodiment of the carbon dioxide monitoring method for an offshore FPSO power station heat pump includes:

[0025] Step S101, performing online component analysis on the flue gas of the FPSO power station to obtain flue gas parameter data, performing heat exchange processing on the flue gas and the flue gas parameter data to obtain pre-processed flue gas and pre-processing parameters;

[0026] Step S102, performing component detection processing on the pre-processed flue gas according to the pre-processing parameters, performing water washing and purification processing on the pre-processed flue gas, and obtaining first clean flue gas and first processing data;

[0027] Step S103, performing temperature field analysis processing on the first clean flue gas based on the first processing data, performing absorption dynamic monitoring processing on the first clean flue gas, and obtaining second clean flue gas, rich amine solution and second processing data;

[0028] Step S104, performing mass transfer detection processing on the second clean flue gas according to the second processing data, performing deep absorption processing on the second clean flue gas, and obtaining third clean flue gas, first lean amine solution and absorption parameters;

[0029] Step S105, performing energy analysis processing on the rich amine liquid based on the absorption parameters, performing heat pump regeneration processing on the rich amine liquid, and obtaining carbon dioxide gas, a second lean amine liquid and regeneration parameters;

[0030] Step S106: Perform purity analysis on the carbon dioxide gas according to the regeneration parameters, compress and liquefy the carbon dioxide gas to obtain liquid carbon dioxide and its quality parameters.

[0031] It is understandable that the execution subject of the present application may be a carbon dioxide monitoring system for an offshore FPSO power station heat pump, or a terminal or a server, which is not specifically limited here. The present application embodiment is described by taking a server as the execution subject as an example.

[0032] Specifically, in the process of CO2 monitoring of heat pumps in offshore FPSO (Floating Production Storage and Offloading) power plants, the processing and analysis of flue gas is a complex and precise task. The flue gas generated during the operation of the FPSO power plant first needs to be analyzed online. The gas chromatography-mass spectrometry combines the high efficiency separation capability of gas chromatography with the precise qualitative and quantitative capability of mass spectrometry, which can quickly and accurately analyze the concentration in flue gas. and other components. During the analysis process, the gas chromatograph first separates the mixed gas into single components, and then the mass spectrometer ionizes these components and detects their mass-to-charge ratio to obtain qualitative and quantitative information of the components. The Fourier transform is used to perform spectrum analysis on the acquired concentration data. The Fourier transform can convert time domain signals into frequency domain signals, making the data easier to analyze and process. For example, when the flue gas is detected When the concentration is 12%, it can be clearly shown by Fourier transform The frequency distribution of characteristic peaks is used to obtain the initial concentration data. Subsequently, the wavelet transform is used to extract dynamic features of the flue gas. Wavelet transform has the characteristics of multi-resolution analysis and can analyze signal characteristics at different scales. The features are screened by an adaptive threshold algorithm to obtain the characteristic parameters of the flue gas. Differential conductivity measurement is used to evaluate the purity of the gas. During the measurement process, a reference conductivity value is set, and the purity is evaluated by measuring the difference between the actual conductivity of the gas and the reference value. The Kalman filter algorithm is used to eliminate measurement noise and improve data reliability. The thermodynamic equilibrium equation is used to calculate the heat transfer, and the nonlinear equation group is solved by the Newton iteration method to obtain the heat transfer control parameters. For example, when the flue gas temperature is 120°C, the optimal heat exchange temperature can be determined to be 80°C through thermodynamic calculations.

[0033] In the water washing purification stage, the ion selective electrode array is used to analyze the ion components. Multiple ion selective electrodes work simultaneously, each electrode specifically detects specific ions, and performs quantitative conversion through Fourier infrared spectroscopy. The multivariate regression algorithm analyzes the correlation between the components, and the support vector regression maps the data to a high-dimensional space for feature extraction. The temperature field analysis uses a distributed optical fiber temperature measurement system, which is laid out in different locations to achieve continuous monitoring of the temperature field. Fast Fourier transform is used to analyze the spectral characteristics of the temperature field and obtain temperature distribution parameters. The concentration detection array is used to monitor the concentration gradient, and the wavelet denoising algorithm eliminates interference signals. The thermocouple array measures the heat change during the absorption process, and the adaptive weighting algorithm fuses the data of multiple measurement points.

[0034] In the mass transfer detection stage, a microfluidic sensor array is used to scan the mass transfer of the interface. Microfluidic technology achieves precise control of gas and liquid phases at the micrometer scale, and wavelet multi-resolution analysis extracts mass transfer characteristics. Fluorescence spectroscopy analysis The migration path and the core principal component algorithm reduce the data dimension. The three-dimensional concentration detection network realizes deep absorption process monitoring, and the genetic algorithm optimizes the absorption parameters. During the heat pump regeneration process, the differential scanning calorimeter determines the heat capacity of the rich amine liquid, and the Raman spectroscopy analyzes the desorption process. The proton transfer reaction mass spectrometry monitors the degradation of the amine liquid, and the thermal resistance temperature sensor array analyzes the temperature field distribution. Finally, the purity of the carbon dioxide gas is analyzed and compressed and liquefied, the gas chromatograph analyzes the impurity content, the quadrupole mass spectrometer performs component analysis, and the pressure sensor array and temperature sensor array monitor the compression and liquefaction process.

[0035] For example, during the initial detection of flue gas from a FPSO power station, gas chromatography-mass spectrometry detected The concentration was 12.5%, and after Fourier transform analysis at 1450 and 2349 Characteristic absorption peaks appeared at the point. After wavelet transform was used to extract dynamic features, the adaptive threshold was set to 0.85, and the main characteristic parameters were screened out. The differential conductivity measurement showed that the purity reached 95%, and stable parameter data was obtained after Kalman filtering. Thermodynamic calculations showed that the optimal heat exchange temperature was 75°C, and the error was less than 0.1% after 5 Newton iterations. During the water washing process, the ion selective electrode detected The concentration is reduced to below 10ppm. The concentration was maintained at 12.3%. Temperature field analysis showed that the temperature gradient in the tower was 3℃ / m, and the mass transfer coefficient reached 0.85mol / ( ·s). Finally, a liquid with a purity of 99.5% was obtained. , pressure is 5.5MPa, temperature is -35℃, meeting the requirements of offshore oil recovery.

[0036] In the embodiment of the present application, by performing online component analysis on the flue gas of the FPSO power station and obtaining flue gas parameter data, real-time monitoring of flue gas components is achieved, the lag of traditional sampling analysis is avoided, and an accurate data basis is provided for subsequent process parameter optimization; by performing heat exchange processing on the flue gas and flue gas parameter data, the waste heat of the system is fully utilized to improve energy utilization efficiency; the pre-treated flue gas is subjected to component detection and water washing purification treatment according to the pre-treatment parameters to effectively remove impurities in the flue gas and create good conditions for the subsequent absorption process; based on the first processing data, the temperature field analysis and absorption dynamic monitoring processing of the first clean flue gas are performed to achieve precise control of the absorption process and improve the absorption efficiency; according to the second processing data, the second clean flue gas is subjected to mass transfer detection and deep absorption processing to ensure the The system can fully capture the carbon dioxide gas and reduce the escape loss. Based on the absorption parameters, the rich amine liquid is subjected to energy analysis and heat pump regeneration treatment to achieve precise control and optimization of the regeneration energy consumption. According to the regeneration parameters, the carbon dioxide gas is subjected to purity analysis and compression liquefaction treatment to obtain a liquid gas that meets the standards. The whole process has established a complete data collection, processing and control system to achieve Intelligent monitoring and control of the entire capture process significantly improves the stability and reliability of the system operation. At the same time, the combined application of multi-stage absorption and heat pump regeneration effectively reduces the system energy consumption and improves economic efficiency. In addition, the solution fully considers the special environment of the offshore platform, adopts distributed monitoring and adaptive control strategies, can adapt to the influence of adverse factors such as equipment shaking, and ensure the continuous and stable operation of the system.

[0037] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0038] (1) The concentration of the flue gas of the FPSO power station is measured by gas chromatography-mass spectrometry to obtain concentration data, and the concentration data is analyzed by Fourier transform to obtain initial concentration data;

[0039] (2) Dynamic characteristics of the FPSO power plant flue gas are extracted and processed by wavelet transform according to the initial concentration data to obtain dynamic characteristics, and the dynamic characteristics are screened and processed by the adaptive threshold algorithm to obtain flue gas characteristic parameters;

[0040] (3) Purity evaluation of the FPSO power plant flue gas is performed through differential conductivity measurement according to the flue gas characteristic parameters to obtain evaluation data, and the evaluation data is subjected to noise elimination processing by the Kalman filter algorithm to obtain flue gas parameter data;

[0041] (4) Based on the flue gas parameter data, the heat exchange of the flue gas is calculated by using the thermodynamic equilibrium equation to obtain the heat exchange data, and the heat exchange data is numerically solved by the Newton iteration method to obtain the heat exchange control parameters;

[0042] (5) According to the heat exchange control parameters, the flue gas is subjected to multi-stage heat exchange processing through temperature gradient analysis to obtain heat exchange data, and the heat exchange data is fitted by the least squares method to obtain the pre-processing temperature curve;

[0043] (6) Based on the pretreatment temperature curve, the heat exchange treatment results are subjected to feature dimension reduction processing by principal component analysis to obtain feature data, and the feature data are classified by a cluster analysis algorithm to obtain the pretreatment flue gas and pretreatment parameters.

[0044] Specifically, in the process of carbon dioxide monitoring of offshore FPSO power station heat pumps, gas chromatography-mass spectrometry is the primary analytical method. Gas chromatography-mass spectrometry combines the separation function of gas chromatography with the qualitative and quantitative functions of mass spectrometry. The gas chromatography part uses a capillary column with an inner diameter of 0.32mm, a coating thickness of 0.25μm, and a carrier gas of helium. The flow rate is controlled at 1.5mL / min. The flue gas sample is injected into the chromatographic column by the injector. Under the programmed temperature condition of 80-200°C, different components are separated in the chromatographic column according to polarity differences. The separated components enter the mass spectrometry ion source, are ionized into charged ions at an electron bombardment energy of 70eV, are separated by a quadrupole mass analyzer, and detected by an electron multiplier to obtain a mass spectrum. Typical The molecular ion peak m / z is 44, and the concentration data is obtained by integrating the peak area. The obtained concentration data is Fourier transformed. Fourier transform is a mathematical tool that converts time domain signals into frequency domain signals. In the digital processing, the fast Fourier transform (FFT) algorithm is used to transform the N-point discrete data sequence to obtain a frequency domain spectrum. The FFT calculation decomposes the original N-point sequence into two N / 2-point sequences, recursively calculates the FFT of each subsequence, and finally merges them to obtain the complete spectrum. By analyzing In 2349 and 667 The characteristic absorption peak intensity at is used to establish a quantitative relationship and obtain the initial concentration data.

[0045] Wavelet transform performs multi-scale analysis on the initial concentration data, and uses the db4 wavelet basis function to perform three-layer decomposition. Each layer of decomposition obtains an approximate coefficient sequence and a detail coefficient sequence. The approximate coefficient reflects the low-frequency trend of the signal, and the detail coefficient reflects the high-frequency details. By reconstructing coefficients of different scales, the dynamic characteristics of flue gas concentration changes are extracted. The adaptive threshold algorithm automatically determines the threshold according to the local statistical characteristics of the signal. When the signal amplitude exceeds the threshold, it is retained, otherwise it is set to zero, thereby screening out significant dynamic characteristics as flue gas characteristic parameters. The differential conductivity measurement adopts the double electrode method, with an electrode spacing of 10mm, an applied AC voltage of 1V, and a frequency of 1kHz. The rate of change of conductivity over time is measured to establish a corresponding relationship with gas purity. The Kalman filter algorithm describes the dynamic characteristics of the system through state equations and observation equations, including two stages: prediction step and correction step. The prediction step predicts the current state based on the state at the previous moment, and the correction step corrects the predicted value based on the observed value, and iterates to obtain smooth flue gas parameter data.

[0046] Thermodynamic equilibrium calculations take into account the physical parameters of the gas mixture, including specific heat capacity, thermal conductivity, etc. Under the condition of pressure P=0.1MPa, the energy conservation equations are established. The Newton iteration method is used to solve the nonlinear equations. After setting the initial value, the function value and Jacobian matrix are calculated. The iterative value is corrected until convergence to obtain the heat exchange data. When the convergence accuracy is set to 0.1%, it usually takes 5-7 iterations to obtain the heat exchange control parameters. Temperature gradient analysis Platinum resistance temperature sensors are arranged at the inlet, middle and outlet of the heat exchanger with a spatial spacing of 0.5m. Record the temperature change data with time and position to construct the temperature field distribution. The least squares method is used to fit the temperature data. The cubic polynomial is selected as the fitting function. The polynomial coefficients are determined by minimizing the residual sum of squares to obtain the preprocessed temperature curve.

[0047] The principal component analysis first standardizes the heat exchange treatment result data and calculates the covariance matrix. The eigenvalue equation is solved to obtain the eigenvalue and eigenvector, and the main eigenvector is selected by sorting the eigenvalue, and the original data is projected into the new feature space to achieve dimensionality reduction. The K-means clustering algorithm classifies the feature data after dimensionality reduction, and sets the number of categories k=3. The center position of the class is optimized through iteration until the difference within the class is the smallest and the difference between the classes is the largest, and the pretreated flue gas and pretreatment parameters are obtained.

[0048] For example, after the flue gas of a FPSO power station was separated by gas chromatography, it was detected at the retention time of 2.3min. The chromatographic peak has a peak area of ​​5680 mV·s. The mass spectrum shows an intensity of 8.2× at m / z 44. The molecular ion peak of The concentration data is 11.5%. FFT analysis shows that at 2349 The amplitude of the absorption peak at is 0.85, corresponding to the fundamental frequency vibration. The three-layer wavelet decomposition obtains 7 wavelet coefficients. After screening by the adaptive threshold (set to 0.75), 3 main coefficients are retained, and the characteristic parameters reflecting the dynamic change of concentration are reconstructed. The differential conductivity measurement shows that the conductivity of pure flue gas is 2.8mS / m, and the measured value is 2.5mS / m. After the Kalman filter eliminates the noise, a stable value of 2.6mS / m is obtained. Thermodynamic calculations show that 4.2kJ / kg of heat needs to be removed under the initial temperature of 115℃. After 6 Newton iterations, the optimal heat transfer temperature is 75℃. The temperature field data shows that the inlet, middle and outlet temperatures are 115℃, 95℃ and 75℃ respectively. The least squares fitting obtains the temperature distribution curve T=115-0.8x (x is the axial distance of the heat exchanger, m). Principal component analysis reduced the original 10-dimensional features to 3 dimensions, retaining 92% of the information. Cluster analysis divided the data into three categories: high, medium, and low, corresponding to different preprocessing parameter intervals.

[0049] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0050] (1) The ion component analysis of the pretreated flue gas is performed by an ion selective electrode array to obtain component data, and the component data is quantitatively converted by Fourier transform infrared spectroscopy to obtain component concentration data;

[0051] (2) Performing correlation analysis on the component concentration data using a multivariate regression algorithm to obtain correlation data, and performing feature mapping on the correlation data using support vector regression to obtain component characteristic parameters;

[0052] (3) The humidity distribution analysis of the component characteristic parameters is performed by water vapor density measurement to obtain distribution data, and the distribution data is optimized by the gradient descent algorithm to obtain the water washing control parameters;

[0053] (4) The pre-treated flue gas is subjected to water washing treatment by countercurrent spraying according to the water washing control parameters to obtain process data, and the process data is subjected to state prediction processing by the Markov chain to obtain purification process data;

[0054] (5) The concentration distribution of the purification process data is analyzed and processed by tomography to obtain distribution data, and the distribution data is dynamically identified and processed by the neural network algorithm to obtain purification effect parameters;

[0055] (6) Performing multi-scale analysis on the purification effect parameters through wavelet packet decomposition to obtain scale data, and performing feature clustering on the scale data through a self-organizing mapping algorithm to obtain a first clean flue gas and a first processed data.

[0056] Specifically, in the carbon dioxide monitoring system of the FPSO power plant heat pump, the ion selective electrode array is an important analytical tool. The array consists of multiple ion selective electrodes, each of which has a selective response to a specific ion. The electrode membrane is made of ion carriers doped in a polyvinyl chloride matrix. When the target ion and the carrier undergo selective complexation, a potential difference signal is generated. Monitoring, using pH electrode and carbonate electrode, the electrode potential and ion concentration are logarithmically related. By real-time acquisition of potential signals, a potential-concentration standard curve is established to obtain component data. Fourier transform infrared spectroscopy analysis uses transmission method, with a scanning range of 400-4000 , resolution 4 , scanning times 32. According to the Beer-Lambert law, absorbance is proportional to concentration, and component concentration data can be obtained by quantitative conversion through integration of characteristic peak areas.

[0057] The multivariate regression algorithm is based on the least squares principle to establish a linear relationship between multiple independent variables and dependent variables. Multivariate regression analysis is performed on the component concentration data, and the interaction between the components is considered to establish a correlation model between the component concentration and the process parameters. Support vector regression maps the data to a high-dimensional space through a kernel function, and finds the optimal separation hyperplane in the high-dimensional space. The radial basis kernel function is used to determine the kernel parameters and penalty factors through cross-validation, and nonlinear mapping and regression analysis are performed on the associated data to obtain the component characteristic parameters. The water vapor density is measured using a dew point meter with a measurement range of -40℃ to 60℃ and an accuracy of ±0.2℃. By measuring the temperature when the flue gas is cooled to the condensation of water vapor, the corresponding saturated water vapor pressure and density are calculated. The water vapor density data at different locations are distributed and analyzed to construct a humidity field distribution model. The gradient descent algorithm finds the minimum value of the objective function in an iterative manner, and updates the parameters along the negative gradient direction each iteration. Set the learning rate to 0.01 and the maximum number of iterations to 1000 times, optimize the distribution data, and obtain the water washing control parameters.

[0058] The countercurrent spray adopts standard nozzle, the working pressure is 0.3MPa, and the liquid-gas ratio is controlled at 1.5L / . The spray liquid is evenly distributed in the form of mist, fully contacts the rising flue gas, and removes the acidic components and particulate matter in the flue gas. The Markov chain describes the state transfer process, establishes the state transfer matrix, and predicts the state probability distribution at the next moment. The purification process data is obtained by analyzing the statistical characteristics of the state sequence. Computer tomography uses a fan beam scanning method with a rotation angle of 360° and an angle interval of 1°. According to the X-ray attenuation law, the projection data is collected and back-projected and reconstructed to obtain a two-dimensional distribution image of the flue gas concentration. The neural network algorithm adopts a three-layer BP network structure, with 20 neurons in the input layer, 30 neurons in the hidden layer, and 10 neurons in the output layer. Using the sigmoid activation function, the network weights are adjusted through back propagation, and pattern recognition is performed on the distribution data to obtain the purification effect parameters.

[0059] Wavelet packet decomposition is an extension of wavelet transform, which decomposes the signal in full frequency band. Using db4 wavelet basis, 4 layers of decomposition are performed to obtain coefficients of 16 frequency bands. Multi-scale features are extracted by analyzing the energy distribution of each frequency band. The self-organizing map algorithm maps high-dimensional data to a two-dimensional plane to maintain the topological relationship of the data. The competition layer neurons are set to a 10×10 rectangular arrangement, and the Gaussian neighborhood function is used to perform cluster analysis on the features through competitive learning to obtain the first clean flue gas and the first processed data.

[0060] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0061] (1) Scanning the temperature field of the first clean flue gas through a distributed optical fiber to obtain temperature field data, and performing spectrum analysis on the temperature field data by fast Fourier transform to obtain temperature distribution parameters;

[0062] (2) By The concentration detection array performs concentration gradient analysis on the temperature distribution parameters to obtain gradient data, and the wavelet denoising algorithm performs noise reduction on the gradient data to obtain concentration field data;

[0063] (3) The concentration field data is analyzed and processed by the thermocouple array to obtain the change data, and the change data is fused by the adaptive weighted algorithm to obtain the absorption thermal parameters;

[0064] (4) Dynamically monitoring the absorption thermal parameters through the liquid level sensor array to obtain monitoring data, and performing parameter estimation processing on the monitoring data by the recursive least squares method to obtain the absorption state parameters;

[0065] (5) Using chemometric methods to perform material balance analysis on the absorption state parameters to obtain equilibrium data, and using particle swarm optimization algorithm to perform parameter optimization on the equilibrium data to obtain an absorption efficiency curve;

[0066] (6) The state data is obtained by performing state recognition processing on the absorption efficiency curve through the fuzzy clustering algorithm, and the state data is processed by probability evaluation through Bayesian inference to obtain the second clean flue gas, rich amine liquid and second treatment data.

[0067] Specifically, in the carbon dioxide monitoring process of the FPSO power plant heat pump, distributed optical fiber temperature field scanning is a key technology. The distributed optical fiber is composed of special quartz optical fiber, with a sensing point set every 1 meter. The total length of the optical fiber is 50 meters and it is arranged around the absorption tower. When the optical signal in the optical fiber encounters temperature changes, a Raman scattering effect will occur, and the temperature information is obtained by analyzing the backscattering spectrum. The data acquisition frequency is 1Hz, the spatial resolution is 0.5 meters, and the temperature measurement accuracy is ±0.1℃. The acquired temperature field data is converted to the frequency domain by fast Fourier transform, the periodic change characteristics of the temperature field are analyzed, the main frequency component is extracted, and the temperature distribution parameters reflecting the temperature distribution law are obtained. The concentration detection array uses the non-dispersive infrared principle and consists of 8 detectors in a circular array, evenly arranged at 45-degree intervals. Each detector contains an infrared light source, an air chamber, a filter, and a thermopile detector. Infrared light is detected in the 2.7μm and 4.3μm bands. Selective absorption, calculate the concentration by measuring the intensity of transmitted light. Set the sampling position according to the temperature distribution parameters to construct the concentration gradient field. The wavelet denoising algorithm uses the db5 wavelet basis function to decompose the gradient data into 5 layers, eliminate noise through threshold processing, and reconstruct the concentration field data.

[0068] The thermocouple array consists of 20 pairs of K-type thermocouples with a temperature measurement range of 0-800℃ and a response time of 0.5 seconds. The heat change is calculated by measuring the temperature difference at different positions, and the thermal effect of the absorption process is analyzed in combination with the concentration field data. The adaptive weighted algorithm dynamically adjusts the weight coefficient according to the reliability of the measurement point, and weightedly fuses the change data of multiple measurement points to obtain the absorption thermal parameters. The liquid level sensor array uses a capacitive level meter with a measurement range of 0-5 meters and an accuracy of ±1mm. 12 sensors are arranged at different heights of the absorption tower to monitor changes in liquid level distribution. The recursive least squares method updates the parameter estimate in real time, tracks the dynamic characteristics of the system, establishes a correlation model between liquid level changes and absorption effects, and obtains the absorption state parameters.

[0069] Chemometric analysis is based on the principles of conservation of mass and energy to establish component balance equations. The material balance relationship is calculated by measuring the flow rate, concentration, temperature and other parameters of the inlet and outlet gas-liquid phases. The particle swarm optimization algorithm sets the population size to 50 and the maximum number of iterations to 1000. The optimal combination of operating parameters is obtained through iterative optimization, and the absorption efficiency curve is drawn. The fuzzy clustering algorithm performs segmented processing on the absorption efficiency curve, sets the membership function and clustering radius, and divides the data points into different state categories. Bayesian inference is based on prior probability and likelihood function, calculates the posterior probability distribution, evaluates the reliability of different states, and finally obtains the second clean flue gas, rich amine liquid and second treatment data.

[0070] For example, the first clean flue gas of a FPSO power station was scanned by distributed optical fiber to obtain data at 50 temperature measurement points, with a temperature range of 40-80°C. Fast Fourier transform shows a significant spectrum peak at 0.05Hz, indicating that the temperature field has a fluctuation period of 20 seconds. The concentration detection array measured the concentrations of 8 points as 11.2%, 11.0%, 10.8%, 10.5%, 10.3%, 10.1%, 9.8%, and 9.6%, respectively, and a radial concentration gradient was constructed. After wavelet denoising, the signal-to-noise ratio was increased from 15dB to 28dB. The thermocouple array measured the heat release of the absorption process as 85kJ / mol, and the standard deviation after adaptive weighted fusion was less than 3%. The liquid level sensor monitored the rich amine layer to be 1.8 meters high, and the absorption coefficient estimated by recursive least squares was 0.82mol / (m³·s). Chemometric analysis showed The removal rate reached 85%, and the particle swarm algorithm optimization obtained the best liquid-gas ratio of 2.5. Fuzzy clustering divided the process into three states: startup, stability, and saturation. Bayesian inference showed that the probability of the stable state was 0.92, and finally obtained The second clean flue gas with a concentration of 8.5% is rich in amine liquid The loading was 0.45 mol / mol.

[0071] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0072] (1) The second clean flue gas is subjected to interface mass transfer scanning processing by a microfluidic sensor array to obtain mass transfer data, and the mass transfer data is subjected to feature extraction processing by wavelet multi-resolution analysis to obtain a mass transfer characteristic spectrum;

[0073] (2) The migration path analysis of the mass transfer characteristic spectrum is performed by fluorescence spectrometer to obtain path data, and the path data is reduced in dimension by the kernel principal component algorithm to obtain migration characteristic parameters;

[0074] (3) The concentration field of the migration characteristic parameters is constructed by the electrochemical sensor array to obtain concentration data, and the concentration data is classified by the random forest algorithm to obtain the absorption area parameters;

[0075] (4) The absorption area parameters are processed in depth through a three-dimensional concentration detection network to obtain absorption data, and the absorption data are optimized by a genetic algorithm to obtain an absorption kinetic curve;

[0076] (5) The absorption kinetics curve is processed by state prediction through a cascade neural network to obtain state data, and the state data is fitted by support vector regression to obtain separation efficiency parameters;

[0077] (6) The separation efficiency parameters are dynamically evaluated and processed by fuzzy time series to obtain evaluation data, and the evaluation data are processed by hierarchical analysis method to obtain the third clean flue gas, the first lean amine liquid and the absorption parameters.

[0078] Specifically, in the carbon dioxide monitoring system of the FPSO power plant heat pump, the microfluidic sensor array consists of 16 integrated microchannels. Each microchannel is 100 microns wide, 50 microns deep, and 10 mm long, and the inner wall of the channel is coated with pH-sensitive fluorescent dye. When mass transfer occurs between the microchannel and the liquid phase, changes in pH value cause changes in fluorescence intensity. Fiber-optic coupling is used to collect fluorescence signals with a scanning frequency of 100 Hz to obtain dynamic data of the mass transfer process. Wavelet multi-resolution analysis uses the Haar wavelet basis function to decompose the mass transfer data into 6 layers, analyze the characteristics at different scales, and reconstruct the mass transfer characteristic spectrum. The fluorescence spectrometer uses an excitation wavelength of 488 nm, a scanning range of 500-600 nm, and a resolution of 0.5 nm. By analyzing the changes in fluorescence intensity over time and space, the mass transfer characteristics of the sample are tracked. The migration path of molecules at the gas-liquid interface. The kernel principal component algorithm uses the Gaussian kernel function for nonlinear mapping, selects the principal component with a cumulative contribution rate of 95%, reduces the high-dimensional path data to a low-dimensional space, and obtains the migration characteristic parameters.

[0079] The electrochemical sensor array consists of 32 microelectrodes with a diameter of 50 microns and a spacing of 200 microns. Concentration distribution, sampling rate 1kHz. The random forest algorithm constructs 100 decision trees, each tree randomly selects some features for training, and classifies the concentration data through a voting mechanism to obtain the absorption area parameters. The three-dimensional concentration detection network arranges 64 detection points in the absorption tower to form a 4×4×4 three-dimensional grid. Each detection point is equipped with a conductivity sensor and a temperature sensor for real-time monitoring. Absorption process. The genetic algorithm sets the chromosome length to 20, the population size to 100, the crossover probability to 0.8, and the mutation probability to 0.1. The absorption parameters are optimized through selection, crossover, and mutation operations to obtain the absorption kinetic curve.

[0080] The cascade neural network consists of three layers of networks connected in series, each layer contains 30, 20, and 10 neurons respectively. The first layer extracts local features, the second layer combines features, and the third layer performs state classification. The network is trained using the back propagation algorithm with a learning rate of 0.01. The support vector regression uses the radial basis kernel function to optimize the kernel parameters and penalty factors through cross-validation, fits the state data, and obtains the separation efficiency parameters. Fuzzy time series analysis considers the temporal characteristics and uncertainty of the data, fuzzifies the separation efficiency parameters, and establishes a fuzzy relationship matrix. The hierarchical analysis method constructs a judgment matrix, calculates the eigenvector, obtains the weights of each evaluation index, comprehensively evaluates the absorption effect, and finally obtains the third clean flue gas, the first lean amine liquid, and the absorption parameters.

[0081] For example: The second clean flue gas of a FPSO power station was detected by a microfluidic sensor array, and the average fluorescence intensity of 16 microchannels was 2850±120 relative units. Wavelet multi-resolution analysis extracted three main characteristic scales, with energy proportions of 65%, 25% and 10% respectively. Fluorescence spectrum analysis showed a characteristic emission peak at 525nm, and the peak intensity decayed exponentially with time, with a half-life of 15 seconds. Nuclear principal component analysis reduced the 128-dimensional spectral data to 5 dimensions, retaining 92% of the information. The electrochemical sensor array measured the local current density range of 0.5-2.5mA / The random forest classification accuracy reached 95%, and the absorption area was divided into three categories: high-efficiency area, transition area and low-efficiency area. The concentration data obtained by the three-dimensional detection network was optimized by genetic algorithm to obtain an absorption rate constant of 0.86 The prediction accuracy of the cascade neural network is 93%, and the root mean square error of the support vector regression fitting is 0.05. The fuzzy time series analysis shows that the system has good dynamic characteristics, and the comprehensive score calculated by the hierarchical analysis method is 0.88. The third clean flue gas with a concentration of 4.5%, the first lean amine liquid The loading was 0.42 mol / mol.

[0082] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0083] (1) The heat capacity of the rich amine liquid is measured by differential scanning calorimetry to obtain heat capacity data, and the heat capacity data is evolved by Fourier heat conduction equation to obtain thermodynamic parameters;

[0084] (2) The heat transfer efficiency is analyzed and processed by the heat flux density sensor array to obtain efficiency data, and the efficiency data is optimized by the steepest descent method to obtain heat transfer state parameters;

[0085] (3) The desorption process of the heat transfer state parameters is analyzed by Raman spectroscopy to obtain desorption data, and the desorption data is separated into components by independent component analysis algorithm to obtain desorption dynamic parameters;

[0086] (4) Degradation monitoring and processing of desorption dynamic parameters are performed by proton transfer reaction mass spectrometry to obtain degradation data, and state estimation processing of the degradation data is performed by the Kalman filter to obtain the regeneration state curve;

[0087] (5) The temperature field distribution of the regeneration state curve is analyzed and processed by the thermal resistance temperature sensor array to obtain temperature data, and the temperature data is optimized by the ant colony optimization algorithm to obtain the regeneration efficiency parameters;

[0088] (6) The regeneration efficiency parameters are comprehensively evaluated and processed by adaptive neuro-fuzzy reasoning to obtain evaluation data, and the evaluation data are classified and processed by a hierarchical clustering algorithm to obtain carbon dioxide gas, the second lean amine liquid and regeneration parameters.

[0089] Specifically, in the carbon dioxide monitoring process of the FPSO power plant heat pump, a differential scanning calorimeter is used to determine the heat capacity of the rich amine liquid. The calorimeter is equipped with a platinum crucible, a sample size of 20 mg, a heating rate of 10°C / min, and a temperature range of 25-200°C. By measuring the heat flow difference between the sample and the reference under the same heating program, a curve of heat capacity versus temperature is obtained. Fourier heat conduction analysis uses a periodic heating method with a heat source frequency of 0.1 Hz. By analyzing the amplitude attenuation and phase delay of the temperature wave, the thermal diffusion coefficient is calculated to obtain the thermodynamic parameters. The heat flux density sensor array consists of 25 heat flux sensors in a 5×5 matrix, each with an area of ​​1 , sensitivity 50μV / (W / ). The sensor is arranged on the surface of the heat exchanger to measure the local heat flux density distribution. The steepest descent method updates the parameters along the negative gradient direction each time through iterative optimization, and the step size is adaptively adjusted to optimize the heat transfer efficiency until convergence, and the heat transfer state parameters are obtained.

[0090] Raman spectroscopy analysis uses 532nm laser excitation, power 100mW, spectral range 200-3600 Through analysis Characteristic peak (1388 ) and amino characteristic peaks (3300 ) intensity ratio to monitor the desorption process. The independent component analysis algorithm assumes that the signal sources are independent of each other. By maximizing the non-Gaussianity of the signal, the mixed spectrum is separated into independent components to obtain the desorption dynamic parameters. Proton transfer reaction mass spectrometry uses As the reactive ion, the reaction temperature is 80℃ and the E / N value is 130Td. The gas phase component changes during the desorption process are monitored in real time to detect degradation products. The Kalman filter establishes a state space model, eliminates measurement noise through prediction-correction iteration, estimates the true state, and obtains the regeneration state curve.

[0091] The thermal resistor temperature sensor array is composed of 36 PT100 platinum resistors with a temperature measurement accuracy of ±0.1°C. A three-dimensional temperature field monitoring network is formed in the regeneration tower. The ant colony optimization algorithm sets the number of ants to 50 and the pheromone volatility coefficient to 0.1. The optimal parameter combination is searched through group intelligence to obtain the regeneration efficiency parameter.

[0092] The adaptive neuro-fuzzy inference system combines the learning ability of neural networks and the reasoning ability of fuzzy logic, and adopts a five-layer network structure. The input layer receives the regeneration efficiency parameters and obtains comprehensive evaluation results through fuzzification, rule reasoning and defuzzification. The hierarchical clustering algorithm adopts the Ward minimum variance method, calculates the distance between classes, and gradually merges clusters from bottom to top to form a hierarchical structure, and obtains carbon dioxide gas, second lean amine liquid and regeneration parameters.

[0093] In a specific embodiment, the process of executing step S106 may specifically include the following steps:

[0094] (1) Perform impurity analysis on carbon dioxide gas using a gas chromatograph to obtain impurity data, and perform feature extraction on the impurity data using a principal component analysis algorithm to obtain purity characteristic parameters;

[0095] (2) The purity characteristic parameters are analyzed by quadrupole mass spectrometry to obtain spectrum data, and the spectrum data is denoised by wavelet transform to obtain component distribution parameters;

[0096] (3) The component distribution parameters are subjected to multi-level compression processing by the pressure sensing array to obtain compressed data, and the compressed data is adjusted by the adaptive control algorithm to obtain the compression state parameters;

[0097] (4) The temperature sensor array is used to analyze the phase change process of the compression state parameters to obtain phase change data, and the neural network algorithm is used to dynamically predict the phase change data to obtain the liquefaction process curve;

[0098] (5) The distributed control system is used to monitor the quality of the liquefaction process curve to obtain monitoring data, and the fuzzy comprehensive evaluation algorithm is used to classify the monitoring data to obtain quality evaluation parameters;

[0099] (6) The quality assessment parameters are comprehensively analyzed and processed by a deep learning algorithm to obtain analysis data, and the analysis data are classified and processed by a decision tree algorithm to obtain liquid carbon dioxide and quality parameters.

[0100] Specifically, in the carbon dioxide monitoring of the FPSO power plant heat pump, the gas chromatograph uses a 30m capillary column for impurity analysis, the column temperature is constant at 80°C, the carrier gas is helium, and the flow rate is 1.5mL / min. The separated components are detected by a flame ionization detector with a sensitivity of g / s. The principal component analysis algorithm was used to process the chromatogram data, calculate the eigenvalues ​​of the covariance matrix, select the eigenvectors with a cumulative contribution rate of 95%, reduce the dimensionality of the multidimensional impurity data, and obtain the purity characteristic parameters. The quadrupole mass spectrometer uses an electron bombardment ion source with an electron energy of 70eV and an ion source temperature of 200°C. The mass scanning range is 1-200amu, and the resolution is 0.1amu. The purity characteristic parameters were analyzed by mass spectrometry to obtain the mass spectrum of the components. The wavelet transform uses the db4 wavelet basis function for 4-layer decomposition, and the baseline drift and random noise are removed by the soft threshold method to reconstruct the component distribution parameters.

[0101] The pressure sensing array contains 16 piezoresistive sensors with a range of 0-10MPa and an accuracy of 0.1%. Sensors are arranged between each stage of the compressor to monitor the pressure distribution. The adaptive control algorithm adjusts the compressor speed and cooling water flow in real time according to the pressure feedback to maintain the optimal compression ratio and obtain the compression state parameters. The temperature sensing array consists of 24 platinum resistance temperature sensors with a temperature measurement range of -50 to 50°C. Monitoring The phase change process from gas to liquid, with a sampling frequency of 10Hz. The neural network adopts LSTM structure, with 64 hidden neurons and 10 time steps, to predict the phase change trend and obtain the liquefaction process curve.

[0102] The distributed control system collects parameters such as pressure, temperature, and flow, with a sampling period of 100ms. The fuzzy comprehensive evaluation establishes an evaluation index system, sets the membership function, calculates the comprehensive score through fuzzy operation, divides the quality into three levels: excellent, good, and qualified, and obtains the quality evaluation parameters. The deep learning algorithm adopts a convolutional neural network structure, which includes 4 convolutional layers and 2 fully connected layers. Input quality evaluation parameters and extract deep features. The decision tree algorithm adopts the CART method, selects the optimal splitting feature through the Gini coefficient, constructs a classification tree, and obtains liquid carbon dioxide and quality parameters.

[0103] For example, the carbon dioxide gas of a FPSO power station was analyzed by gas chromatography, and trace moisture 150ppm, oxygen 80ppm, and carbon monoxide 35ppm were detected. Principal component analysis extracted three principal components, with a cumulative variance contribution rate of 96%. The quadrupole mass spectrometer detected at m / z44 The main peak has a relative abundance of 100%, and the isotope peak at m / z 45 has a relative abundance of 1.1%. After wavelet denoising, the signal-to-noise ratio is improved to 35dB. The pressure sensor array shows that the four-stage compression pressure is 0.6, 1.8, 3.5, and 6.2MPa respectively. The adaptive control stabilizes the compressor speed at 1450rpm. The temperature sensor monitors that at 5.8MPa The liquefaction started, the temperature dropped to -18°C, and the neural network predicted the final temperature to be -35°C. The fuzzy judgment comprehensive score was 0.89, which was of excellent quality. Deep learning extracted 12 key features, and the decision tree divided the products into high-purity and industrial grades, and finally obtained liquid carbon dioxide with a purity of 99.95%, a pressure of 6.2MPa, and a temperature of -35°C.

[0104] The above describes the carbon dioxide monitoring method for the offshore FPSO power plant heat pump in the embodiment of the present application. The following describes the carbon dioxide monitoring system for the offshore FPSO power plant heat pump in the embodiment of the present application. Figure 2 In the embodiment of the present application, an embodiment of a carbon dioxide monitoring system for an offshore FPSO power station heat pump includes:

[0105] The analysis module 201 is used to perform online component analysis on the flue gas of the FPSO power station, obtain flue gas parameter data, perform heat exchange processing on the flue gas and the flue gas parameter data, and obtain pre-processed flue gas and pre-processing parameters;

[0106] A detection module 202, configured to perform component detection processing on the pre-treated flue gas according to the pre-treatment parameters, and perform water washing purification processing on the pre-treated flue gas to obtain first clean flue gas and first processing data;

[0107] A processing module 203 is used to perform temperature field analysis processing on the first clean flue gas based on the first processing data, and perform absorption dynamic monitoring processing on the first clean flue gas to obtain second clean flue gas, rich amine liquid and second processing data;

[0108] an absorption module 204, configured to perform mass transfer detection processing on the second clean flue gas according to the second processing data, and perform deep absorption processing on the second clean flue gas to obtain third clean flue gas, first lean amine liquid and absorption parameters;

[0109] A regeneration module 205, for performing energy analysis processing on the rich amine liquid based on the absorption parameters, performing heat pump regeneration processing on the rich amine liquid, and obtaining carbon dioxide gas, a second lean amine liquid and regeneration parameters;

[0110] The liquefaction module 206 is used to perform purity analysis on the carbon dioxide gas according to the regeneration parameters, and to compress and liquefy the carbon dioxide gas to obtain liquid carbon dioxide and its quality parameters.

[0111] Through the coordinated cooperation of the above-mentioned components, by conducting online component analysis and processing of the FPSO power plant flue gas and obtaining flue gas parameter data, real-time monitoring of flue gas components is achieved, avoiding the lag of traditional sampling analysis, and providing an accurate data basis for subsequent process parameter optimization; by performing heat exchange processing on the flue gas and flue gas parameter data, the system waste heat is fully utilized to improve energy utilization efficiency; the pre-treated flue gas is subjected to component detection and water washing purification according to the pre-treatment parameters to effectively remove impurities in the flue gas and create good conditions for the subsequent absorption process; based on the first processing data, the first clean flue gas is subjected to temperature field analysis and absorption dynamic monitoring processing to achieve precise control of the absorption process and improve the absorption efficiency; the second clean flue gas is subjected to mass transfer detection and deep absorption processing according to the second processing data to ensure through multi-stage absorption The system can fully capture the carbon dioxide gas and reduce the escape loss. Based on the absorption parameters, the rich amine liquid is subjected to energy analysis and heat pump regeneration treatment to achieve precise control and optimization of the regeneration energy consumption. According to the regeneration parameters, the carbon dioxide gas is subjected to purity analysis and compression liquefaction treatment to obtain a liquid gas that meets the standards. The whole process has established a complete data collection, processing and control system to achieve Intelligent monitoring and control of the entire capture process significantly improves the stability and reliability of the system operation. At the same time, the combined application of multi-stage absorption and heat pump regeneration effectively reduces the system energy consumption and improves economic efficiency. In addition, the solution fully considers the special environment of the offshore platform, adopts distributed monitoring and adaptive control strategies, can adapt to the influence of adverse factors such as equipment shaking, and ensure the continuous and stable operation of the system.

[0112] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0113] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A carbon dioxide monitoring method for an offshore FPSO power station heat pump, characterized in that: The carbon dioxide monitoring method for an offshore FPSO power station heat pump comprises: Performing online component analysis on the flue gas of the FPSO power station to obtain flue gas parameter data, performing heat exchange processing on the flue gas and the flue gas parameter data to obtain pre-treated flue gas and pre-treatment parameters, including: performing concentration measurement on the flue gas of the FPSO power station by gas chromatography-mass spectrometry to obtain concentration data, performing spectrum analysis on the concentration data by Fourier transform to obtain initial concentration data; performing dynamic feature extraction on the flue gas of the FPSO power station by wavelet transform according to the initial concentration data to obtain dynamic features, performing screening on the dynamic features by an adaptive threshold algorithm to obtain flue gas characteristic parameters; performing purity evaluation on the flue gas of the FPSO power station by differential conductivity measurement according to the flue gas characteristic parameters to obtain evaluation data, performing noise elimination on the evaluation data by a Kalman filter algorithm to obtain the flue gas parameter data; Based on the flue gas parameter data, the heat exchange calculation processing of the flue gas is performed through the thermodynamic equilibrium equation to obtain heat exchange data, and the heat exchange data is numerically solved by the Newton iteration method to obtain heat exchange control parameters; according to the heat exchange control parameters, the flue gas is subjected to multi-stage heat exchange processing through temperature gradient analysis to obtain heat exchange data, and the heat exchange data is fitted by the least square method to obtain a pretreatment temperature curve; based on the pretreatment temperature curve, the heat exchange processing result is subjected to feature dimension reduction processing through principal component analysis to obtain feature data, and the feature data is classified by a cluster analysis algorithm to obtain the pretreated flue gas and the pretreatment parameters; The pre-treated flue gas is subjected to component detection processing according to the pre-treatment parameters, and the pre-treated flue gas is subjected to water washing and purification processing to obtain a first clean flue gas and a first processing data, including: performing ion component analysis processing on the pre-treated flue gas through an ion selective electrode array to obtain component data, performing quantitative conversion processing on the component data through Fourier transform infrared spectroscopy to obtain component concentration data; performing correlation analysis processing on the component concentration data through a multivariate regression algorithm to obtain correlation data, performing feature mapping processing on the correlation data through support vector regression to obtain component characteristic parameters; performing humidity distribution analysis processing on the component characteristic parameters through water vapor density measurement to obtain distribution data, and performing gradient spectroscopy to obtain component concentration data; performing correlation analysis processing on the component concentration data through a multivariate regression algorithm to obtain correlation data, and ... The distribution data is optimized by a descent algorithm to obtain a water washing control parameter; the pre-treated flue gas is subjected to water washing by countercurrent spraying according to the water washing control parameter to obtain process data, and the process data is subjected to state prediction processing by a Markov chain to obtain purification process data; the purification process data is subjected to concentration distribution analysis processing by tomography to obtain distribution data, and the distribution data is subjected to dynamic identification processing by a neural network algorithm to obtain purification effect parameters; the purification effect parameters are subjected to multi-scale analysis processing by wavelet packet decomposition to obtain scale data, and the scale data is subjected to feature clustering processing by a self-organizing map algorithm to obtain the first clean flue gas and the first processed data; Based on the first processed data, the first clean flue gas is subjected to temperature field analysis and processing, and the first clean flue gas is subjected to absorption dynamic monitoring processing to obtain second clean flue gas, rich amine liquid and second processed data, including: performing temperature field scanning processing on the first clean flue gas through distributed optical fiber to obtain temperature field data, performing spectrum analysis processing on the temperature field data through fast Fourier transform to obtain temperature distribution parameters; performing concentration gradient analysis processing on the temperature distribution parameters through concentration detection array to obtain gradient data, performing denoising processing on the gradient data through wavelet denoising algorithm to obtain concentration field data; performing heat change analysis processing on the concentration field data through thermocouple array to obtain change data, and performing adaptive heating The weight algorithm is used to fuse the change data to obtain absorption thermal parameters; the absorption thermal parameters are dynamically monitored by a liquid level sensor array to obtain monitoring data, and the monitoring data are estimated by a recursive least squares method to obtain absorption state parameters; the absorption state parameters are subjected to material balance analysis by a chemometric method to obtain balance data, and the balance data are subjected to parameter optimization by a particle swarm optimization algorithm to obtain an absorption efficiency curve; the absorption efficiency curve is subjected to state recognition by a fuzzy clustering algorithm to obtain state data, and the state data are subjected to probability evaluation by Bayesian inference to obtain the second clean flue gas, the rich amine liquid and the second processed data; Performing mass transfer detection processing on the second clean flue gas according to the second processing data, and performing deep absorption processing on the second clean flue gas to obtain third clean flue gas, first lean amine liquid and absorption parameters; Performing energy analysis processing on the rich amine liquid based on the absorption parameters, and performing heat pump regeneration processing on the rich amine liquid to obtain carbon dioxide gas, a second lean amine liquid and regeneration parameters; The carbon dioxide gas is subjected to purity analysis according to the regeneration parameters, and the carbon dioxide gas is subjected to compression and liquefaction processing to obtain liquid carbon dioxide and its quality parameters.

2. The carbon dioxide monitoring method for an offshore FPSO power station heat pump according to claim 1, characterized in that: The method of performing mass transfer detection processing on the second clean flue gas according to the second processing data and performing deep absorption processing on the second clean flue gas to obtain third clean flue gas, first lean amine liquid and absorption parameters includes: Performing interface mass transfer scanning processing on the second clean flue gas by a microfluidic sensor array to obtain mass transfer data, and performing feature extraction processing on the mass transfer data by wavelet multi-resolution analysis to obtain a mass transfer characteristic spectrum; Performing migration path analysis on the mass transfer characteristic spectrum by fluorescence spectrometer to obtain path data, and performing dimension reduction processing on the path data by kernel principal component algorithm to obtain migration characteristic parameters; Performing concentration field construction processing on the migration characteristic parameters through an electrochemical sensor array to obtain concentration data, and performing classification processing on the concentration data through a random forest algorithm to obtain absorption region parameters; Performing deep absorption processing on the absorption region parameters through a three-dimensional concentration detection network to obtain absorption data, and performing optimization processing on the absorption data through a genetic algorithm to obtain an absorption kinetics curve; Performing state prediction processing on the absorption kinetics curve by a cascade neural network to obtain state data, and performing fitting processing on the state data by support vector regression to obtain separation efficiency parameters; The separation efficiency parameters are dynamically evaluated and processed by fuzzy time series to obtain evaluation data, and the evaluation data are subject to decision-making processing by hierarchical analysis to obtain the third clean flue gas, the first lean amine liquid and the absorption parameters.

3. The carbon dioxide monitoring method for an offshore FPSO power station heat pump according to claim 1, characterized in that: The step of performing energy analysis on the rich amine liquid based on the absorption parameters and performing heat pump regeneration on the rich amine liquid to obtain carbon dioxide gas, a second lean amine liquid and regeneration parameters comprises: The rich amine liquid is subjected to heat capacity measurement by a differential scanning calorimeter to obtain heat capacity data, and the heat capacity data is subjected to evolution processing by a Fourier heat conduction equation to obtain thermodynamic parameters; Performing heat exchange efficiency analysis on the thermodynamic parameters by means of a heat flux density sensor array to obtain efficiency data, and optimizing the efficiency data by means of a steepest descent method to obtain heat exchange state parameters; Performing desorption process analysis on the heat exchange state parameters by Raman spectroscopy to obtain desorption data, and performing component separation on the desorption data by independent component analysis algorithm to obtain desorption dynamic parameters; Performing degradation monitoring processing on the desorption dynamic parameters by proton transfer reaction mass spectrometry to obtain degradation data, and performing state estimation processing on the degradation data by a Kalman filter to obtain a regeneration state curve; Performing temperature field distribution analysis on the regeneration state curve by using a thermal resistance temperature sensor array to obtain temperature data, and performing parameter optimization processing on the temperature data by using an ant colony optimization algorithm to obtain a regeneration efficiency parameter; The regeneration efficiency parameters are comprehensively evaluated and processed by adaptive neural fuzzy reasoning to obtain evaluation data, and the evaluation data are classified and processed by a hierarchical clustering algorithm to obtain the carbon dioxide gas, the second lean amine liquid and the regeneration parameters.

4. The carbon dioxide monitoring method for an offshore FPSO power station heat pump according to claim 1, characterized in that: The method of performing purity analysis on the carbon dioxide gas according to the regeneration parameters, compressing and liquefying the carbon dioxide gas to obtain liquid carbon dioxide and its quality parameters includes: Performing impurity analysis on the carbon dioxide gas through a gas chromatograph to obtain impurity data, and performing feature extraction on the impurity data through a principal component analysis algorithm to obtain purity characteristic parameters; Performing component spectrum analysis on the purity characteristic parameters by quadrupole mass spectrometry to obtain spectrum data, and performing denoising on the spectrum data by wavelet transform to obtain component distribution parameters; Performing multi-level compression processing on the component distribution parameters through a pressure sensing array to obtain compressed data, and performing adjustment processing on the compressed data through an adaptive control algorithm to obtain compression state parameters; Performing phase change process analysis on the compression state parameters by a temperature sensor array to obtain phase change data, and performing dynamic prediction processing on the phase change data by a neural network algorithm to obtain a liquefaction process curve; Performing quality monitoring processing on the liquefaction process curve by means of a distributed control system to obtain monitoring data, and performing grade classification processing on the monitoring data by means of a fuzzy comprehensive evaluation algorithm to obtain quality evaluation parameters; The quality assessment parameters are comprehensively analyzed and processed by a deep learning algorithm to obtain analysis data, and the analysis data are classified and processed by a decision tree algorithm to obtain the liquid carbon dioxide and the quality parameters.

5. A carbon dioxide monitoring system for an offshore FPSO power station heat pump, used to implement the carbon dioxide monitoring method for an offshore FPSO power station heat pump as described in any one of claims 1 to 4, characterized in that: The carbon dioxide monitoring system for an offshore FPSO power plant heat pump comprises: The analysis module is used to perform online component analysis on the flue gas of the FPSO power station, obtain flue gas parameter data, perform heat exchange processing on the flue gas and the flue gas parameter data to obtain pre-treated flue gas and pre-treatment parameters, perform concentration measurement processing on the flue gas of the FPSO power station by gas chromatography-mass spectrometry to obtain concentration data, perform spectrum analysis processing on the concentration data by Fourier transform to obtain initial concentration data; perform dynamic feature extraction processing on the flue gas of the FPSO power station by wavelet transform according to the initial concentration data to obtain dynamic features, perform screening processing on the dynamic features by an adaptive threshold algorithm to obtain flue gas characteristic parameters; perform purity evaluation processing on the flue gas of the FPSO power station by differential conductivity measurement according to the flue gas characteristic parameters to obtain evaluation data, performing noise elimination processing on the evaluation data by a Kalman filter algorithm to obtain the flue gas parameter data; performing heat exchange calculation processing on the flue gas by a thermodynamic equilibrium equation based on the flue gas parameter data to obtain heat exchange data, performing numerical solution processing on the heat exchange data by a Newton iteration method to obtain heat exchange control parameters; performing multi-stage heat exchange processing on the flue gas by temperature gradient analysis according to the heat exchange control parameters to obtain heat exchange data, performing fitting processing on the heat exchange data by a least square method to obtain a pretreatment temperature curve; performing feature dimension reduction processing on the heat exchange processing result by principal component analysis based on the pretreatment temperature curve to obtain feature data, performing classification processing on the feature data by a cluster analysis algorithm to obtain the pretreated flue gas and the pretreatment parameters; The detection module is used to perform component detection processing on the pre-treated flue gas according to the pre-treatment parameters, and perform water washing and purification processing on the pre-treated flue gas to obtain first clean flue gas and first processing data, including: performing ion component analysis processing on the pre-treated flue gas through an ion selective electrode array to obtain component data, and performing quantitative conversion processing on the component data through Fourier infrared spectroscopy to obtain component concentration data; performing correlation analysis processing on the component concentration data through a multivariate regression algorithm to obtain correlation data, and performing feature mapping processing on the correlation data through support vector regression to obtain component characteristic parameters; performing humidity distribution analysis processing on the component characteristic parameters through water vapor density measurement to obtain distribution data, The distribution data is optimized by a gradient descent algorithm to obtain a water washing control parameter; the pre-treated flue gas is subjected to water washing by countercurrent spraying according to the water washing control parameter to obtain process data, and the process data is subjected to state prediction processing by a Markov chain to obtain purification process data; the purification process data is subjected to concentration distribution analysis processing by tomography to obtain distribution data, and the distribution data is subjected to dynamic identification processing by a neural network algorithm to obtain purification effect parameters; the purification effect parameters are subjected to multi-scale analysis processing by wavelet packet decomposition to obtain scale data, and the scale data is subjected to feature clustering processing by a self-organizing map algorithm to obtain the first clean flue gas and the first processed data; a processing module, for performing temperature field analysis processing on the first clean flue gas based on the first processing data, performing absorption dynamic monitoring processing on the first clean flue gas, and obtaining second clean flue gas, rich amine solution, and second processing data, including: performing temperature field scanning processing on the first clean flue gas through distributed optical fiber to obtain temperature field data, performing spectrum analysis processing on the temperature field data through fast Fourier transform to obtain temperature distribution parameters; performing concentration gradient analysis processing on the temperature distribution parameters through a concentration detection array to obtain gradient data, performing denoising processing on the gradient data through a wavelet denoising algorithm to obtain concentration field data; performing heat change analysis processing on the concentration field data through a thermocouple array to obtain change data, and ... and performing denoising processing on the gradient data through a wavelet denoising algorithm to obtain concentration field data; performing heat change analysis processing on the concentration field data through a thermocouple array to obtain change data, and performing spectrum analysis processing on the temperature field data through fast Fourier transform to obtain temperature distribution parameters; An adaptive weighted algorithm is used to fuse the change data to obtain absorption thermal parameters; the absorption thermal parameters are dynamically monitored by a liquid level sensor array to obtain monitoring data, and the monitoring data are estimated by a recursive least squares method to obtain absorption state parameters; the absorption state parameters are subjected to material balance analysis by a chemometric method to obtain balance data, and the balance data are subjected to parameter optimization by a particle swarm optimization algorithm to obtain an absorption efficiency curve; the absorption efficiency curve is subjected to state recognition by a fuzzy clustering algorithm to obtain state data, and the state data are subjected to probability evaluation by Bayesian inference to obtain the second clean flue gas, the rich amine liquid and the second processed data; an absorption module, configured to perform mass transfer detection processing on the second clean flue gas according to the second processing data, perform deep absorption processing on the second clean flue gas, and obtain third clean flue gas, first lean amine liquid and absorption parameters; a regeneration module, for performing energy analysis processing on the rich amine liquid based on the absorption parameters, performing heat pump regeneration processing on the rich amine liquid, and obtaining carbon dioxide gas, a second lean amine liquid and regeneration parameters; The liquefaction module is used to perform purity analysis on the carbon dioxide gas according to the regeneration parameters, and to compress and liquefy the carbon dioxide gas to obtain liquid carbon dioxide and its quality parameters.

Images