Intelligent control method and system for real-time treatment of marine exhaust gas

Abstract

The application discloses an intelligent control method and system for real-time treatment of marine exhaust gas, which comprises the following steps: collecting the concentration signal of the exhaust gas components and the dynamic parameters of the sea conditions, extracting the vortex intensity of the exhaust gas component concentration signal to identify the vortex flow characteristics, extracting the flow field disturbance coefficient from the dynamic parameters of the sea conditions, and constructing the exhaust gas-environment interaction field; based on the interaction field, tracking and positioning the high-temperature pollution source by heat trace, constructing the heat propagation path and mapping the exhaust gas heat flow network; performing turbulence analysis on the heat flow network to identify the vortex breaking point, determining the best treatment position and calibrating the treatment axis; based on the treatment axis, recombining the control parameter chain, determining the steady-state treatment interval and generating the adaptive adjustment matrix; fusing the adjustment matrix and the temperature gradient data to identify the latent heat region of phase change, extracting the condensation release energy to construct the energy recovery tensor, and generating the cascade treatment scheme; forming a feedback curve by monitoring the execution deviation, generating an exhaust gas treatment execution instruction, and ensuring the continuous and stable operation of the marine exhaust gas treatment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of marine environmental protection, in particular to an intelligent control method and system for real-time treatment of marine exhaust gas. BACKGROUND

[0002] Ship exhaust gas emission has become an important source of marine environmental pollution, and existing exhaust gas treatment technology has many limitations in dealing with complex marine environments. The traditional exhaust gas diffusion analysis method mainly relies on a simplified diffusion model, which is difficult to accurately describe the real propagation behavior of exhaust gas under variable sea conditions, resulting in a lack of scientificity in the arrangement of treatment equipment.

[0003] Existing exhaust gas treatment control technology is usually based on fixed parameter settings and lacks the ability to respond to real-time changes in the marine environment, and cannot dynamically adjust according to sea conditions and exhaust gas characteristics. In terms of energy utilization, the traditional method is mainly designed for fixed working conditions, and the heat energy resources in the exhaust gas have not been fully tapped, and the energy recovery efficiency needs to be improved. Therefore, there is an urgent need for a method to solve at least one of the above problems. SUMMARY

[0004] The present application discloses an intelligent control method and system for real-time treatment of marine exhaust gas, aiming to build an exhaust gas-environment interaction field, deeply analyze the diffusion and transmission mechanism of exhaust gas in the marine environment, and realize intelligent identification, accurate treatment and efficient energy recovery of ship exhaust gas by combining vortex feature recognition, heat trace tracking positioning, latent heat recovery and other technologies, forming a self-adaptive hierarchical treatment scheme and real-time control strategy.

[0005] The present application discloses an intelligent control method and system for real-time treatment of marine exhaust gas, aiming to build an exhaust gas-environment interaction field, deeply analyze the diffusion and transmission mechanism of exhaust gas in the marine environment, and realize intelligent identification, accurate treatment and efficient energy recovery of ship exhaust gas by combining vortex feature recognition, heat trace tracking positioning, latent heat recovery and other technologies, forming a self-adaptive hierarchical treatment scheme and real-time control strategy.

[0006] Collecting ship exhaust gas composition concentration signals and sea condition dynamic parameters, identifying vortex flow characteristics by vortex intensity extraction on the exhaust gas composition concentration signals, extracting flow field disturbance coefficients from the sea condition dynamic parameters, and constructing an exhaust gas-environment interaction field by time series correlation of the vortex flow characteristics and the flow field disturbance coefficients;

[0007] Based on the exhaust gas-environment interaction field, heat trace tracking positioning of high-temperature pollution sources is performed, temperature field diffusion analysis is performed on the high-temperature pollution sources to construct a heat propagation path, temperature gradient data is collected along the heat propagation path, and the temperature gradient data is used to map an exhaust gas heat flow network;

[0008] Performing turbulent analysis on the exhaust gas heat flow network to identify vortex breaking points, extracting mixing intensity parameters of the vortex breaking points, determining the best treatment position according to the correlation degree of the mixing intensity parameters and the vortex flow characteristics, and connecting the best treatment position with the high-temperature pollution source to calibrate a treatment axis;

[0009] reorganize control parameter chains based on the processing axis, match the control parameter chains with the flow field disturbance coefficient to determine a steady-state processing interval, and generate an adaptive adjustment matrix using the steady-state processing interval;

[0010] fuse the adaptive adjustment matrix with the temperature gradient data to identify a latent heat of phase change area, extract condensation release energy in the latent heat of phase change area, construct an energy recovery tensor based on an energy difference between the condensation release energy and the waste gas heat flow network, and generate a cascade processing scheme through the energy recovery tensor;

[0011] monitor execution deviations of the cascade processing scheme to form a feedback curve, generate waste gas processing execution instructions based on the feedback curve, and complete real-time processing control of marine waste gas.

[0012] The second aspect of the present application proposes an intelligent control system for real-time processing of marine waste gas, comprising:

[0013] A data acquisition module is configured to acquire a ship waste gas component concentration signal and a sea state dynamic parameter, extract a vortex intensity from the waste gas component concentration signal to identify a vortex flow feature, extract a flow field disturbance coefficient from the sea state dynamic parameter, and construct a waste gas-environment interaction field by time-series correlation between the vortex flow feature and the flow field disturbance coefficient.

[0014] A heat flow analysis module is configured to perform heat tracking on the waste gas-environment interaction field to locate a high-temperature pollution source, perform temperature field diffusion analysis on the high-temperature pollution source to construct a heat propagation path, and acquire temperature gradient data along the heat propagation path to map a waste gas heat flow network using the temperature gradient data.

[0015] A turbulent flow positioning module is configured to perform turbulent flow analysis on the waste gas heat flow network to identify a vortex breaking point, extract a mixing intensity parameter of the vortex breaking point, determine an optimal processing location according to a correlation degree between the mixing intensity parameter and the vortex flow feature, and label a processing axis by connecting the optimal processing location with the high-temperature pollution source.

[0016] A parameter configuration module is configured to reorganize control parameter chains based on the processing axis, match the control parameter chains with the flow field disturbance coefficient to determine a steady-state processing interval, and generate an adaptive adjustment matrix using the steady-state processing interval.

[0017] An energy recovery module is configured to fuse the adaptive adjustment matrix with the temperature gradient data to identify a latent heat of phase change area, extract condensation release energy in the latent heat of phase change area, construct an energy recovery tensor based on an energy difference between the condensation release energy and the waste gas heat flow network, and generate a cascade processing scheme through the energy recovery tensor.

[0018] The execution control module is used for monitoring execution deviation of the step treatment scheme to form a feedback curve, generating waste gas treatment execution instructions based on the feedback curve, and completing real-time treatment control of marine waste gas.

[0019] The beneficial effects of the present application are embodied in the following points: first, by vortex intensity extraction of waste gas component concentration signals and flow field disturbance coefficient acquisition of sea state dynamic parameters, an exhaust gas-environment interaction field is constructed, the quantitative correlation of exhaust gas diffusion behavior and marine environmental conditions is realized, the accuracy of exhaust gas transmission path prediction is improved, and the spatial precise positioning of exhaust gas treatment is realized through accurate positioning of high-temperature pollution source position by heat trace tracking technology. Secondly, the vortex breaking point is identified by turbulent flow analysis and the mixing intensity parameter is extracted, the adaptive adjustment matrix is constructed by combining the resonant frequency scanning and phase locking technology, the dynamic optimization configuration of the treatment parameter is realized, the exhaust gas treatment efficiency and system stability are improved, and the reliable operation of the control system in complex marine environment is ensured through the determination of the steady-state treatment interval. Finally, through phase change latent heat zone identification and condensation release energy extraction, an energy recovery tensor is constructed and a step treatment scheme is generated, the graded recycling of exhaust gas heat energy is realized, the overall energy consumption of the system is reduced, and a complete real-time control closed loop is formed through execution deviation monitoring and feedback curve generation, ensuring the continuous and stable operation of marine waste gas treatment.

[0020] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. BRIEF DESCRIPTION OF DRAWINGS

[0021] The drawings herein show specific examples of the technical solutions described in the present application, and constitute part of the specification together with the specific embodiments, for explaining the technical solutions, principles and effects of the present application.

[0022] Unless specifically stated or otherwise, the same reference signs in different drawings represent the same or similar technical features, and different reference signs may also be used to represent the same or similar technical features.

[0023] Figure 1 is a flow diagram of an intelligent control method for real-time treatment of marine waste gas according to the present application.

[0024] Figure 2 is a structural block diagram of an intelligent control system for real-time treatment of marine waste gas according to the present application. DETAILED DESCRIPTION

[0025] In the following description, for purposes of explanation and not limitation, specific details are set forth such as particular architectures, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known methods, devices, circuits, and

[0026] It is to be understood that the terminology "including", "comprising", "consisting" and "consisting essentially of" used in the specification and the appended claims, indicates the presence of the stated features, integers, steps, operations, elements, and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0027] Reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the application. Thus, the appearances of the phrases "in one embodiment" or "in an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment, but can refer to one or more but not all embodiments. Furthermore, the terms "a" or "an", as used herein, mean "one or more" unless otherwise explicitly stated.

[0028] The technical solutions of the embodiments of the present application are introduced as follows.

[0029] As shown in Figure 1 The embodiment of the present application provides an intelligent control method for real-time treatment of marine exhaust gas, which comprises the following steps S110-S160:

[0030] In step S110, the concentration signals of the ship exhaust gas components and the sea state dynamic parameters are collected, the vortex intensity of the exhaust gas component concentration signals is extracted to identify the vortex flow characteristics, the flow field disturbance coefficient is extracted from the sea state dynamic parameters, and the vortex flow characteristics and the flow field disturbance coefficient are time-sequentially associated to construct an exhaust gas-environment interaction field.

[0031] Specifically, the ship exhaust composition concentration signal and the sea state dynamic parameters are collected. The concentration changes of the main pollutants in the exhaust composition concentration signal are continuously monitored through the gas concentration sensor array arranged at the ship chimney outlet and the surrounding area. The sensor array includes sulfur dioxide sensors, nitrogen oxide sensors, particulate matter concentration detectors, and carbon monoxide sensors, which realize high-precision measurement by using electrochemical principle and light scattering technology. The sensors are installed in a grid layout, covering the chimney outlet and its downwind area, to ensure the capture of transient concentration fluctuations in the diffusion process of the exhaust composition concentration signal. The data collection adopts a high-frequency continuous recording mode to ensure that the time resolution meets the requirements of turbulence analysis. The sea state dynamic parameters, including sea surface wind speed, wind direction, wave height, wave period, sea water temperature, and atmospheric pressure, are collected synchronously. The sea state dynamic parameters are obtained through the shipborne weather station and the wave radar. The weather station is equipped with an ultrasonic wind speed and direction instrument, a barometric pressure sensor, and a temperature and humidity sensor. The wave radar uses X-band radar technology to measure the sea surface wave spectrum and effective wave height in real time. All sensor data are transmitted to the central data collection system through the bus. The collection system is equipped with a high-speed data cache and a time synchronization module to ensure the consistency of the time stamps and the spatial correlation of the exhaust composition concentration signal and the sea state dynamic parameters.

[0032] In some embodiments, the vortex intensity extraction of the exhaust composition concentration signal identifies the vortex flow feature, including: constructing a vortex evolution graph based on the exhaust composition concentration signal; obtaining a vortex core origin point by reverse tracking processing the vortex evolution graph; identifying a circulation path by injecting a tracer disturbance into the vortex core origin point; and extracting a rotation parameter along the circulation path to encode as a vortex flow feature.

[0033] A vortex evolution graph is constructed based on the exhaust composition concentration signal. The time series data of the exhaust composition concentration signal is organized into a four-dimensional data cube according to the spatial position of the sensor and the time dimension, establishing a spatiotemporal distribution model of the concentration signal field. The spatial gradient and time derivative of the exhaust composition concentration signal field are calculated using the finite difference method, and the flow field structure is inversely calculated from the concentration signal distribution by solving the inverse problem of the concentration transport equation. The velocity field is separated into irrotational potential flow and rotational vortex flow using flow field decomposition technology, and the rotational component is extracted to construct the vorticity field distribution. The vorticity where ω is the vorticity vector, is the gradient operator, and v is the velocity field vector. The contour lines of the vorticity field describe the spatial morphology of the vortex structure, and the vortex evolution graph is established by connecting the vorticity distributions at different times. The horizontal axis of the graph represents time, the vertical axis represents spatial position, and the color depth represents vorticity intensity. The vortex trajectory in the vortex evolution graph shows the spatiotemporal evolution process of the vortex core, including the complete life cycle of vortex generation, development, merging, and dissipation. Wavelet transform is used to perform multi-scale decomposition on the graph to identify vortex motion patterns at different time scales and distinguish between large-scale organizational structures and small-scale turbulent vortices.

[0034] The vortex evolution map is processed to obtain the vortex core origin point. The local maximum points of vorticity intensity are identified in the vortex evolution map, which correspond to the vortex core positions. The reverse time integration method is used to track the historical trajectory of the vortex core, and the position of the vortex core origin point is determined by solving the particle motion equation in reverse. The reverse tracking uses a high-precision numerical integration format to ensure tracking accuracy and numerical stability. A decay model of vortex core intensity over time is established, and the decay law and model parameters are determined by fitting the observed data in the vortex evolution map. When the vortex core intensity is lower than the background noise level, the position is determined as the vortex core origin point. Multiple vortex cores in the map are simultaneously processed by reverse tracking to identify the spatial distribution pattern and statistical characteristics of the vortex core origin point. The analysis of the origin point reveals the preferred location and generation mechanism of the vortex, including shear layer instability, boundary layer separation, and obstacle flow around the physical process. The spatiotemporal coordinates, initial intensity, and generation conditions of each vortex core origin point are recorded to establish an origin point database, providing basic data for vortex prediction and control. The vortex core origin points are classified by clustering analysis method to identify different types of vortex generation patterns and establish a classification system of vortex generation.

[0035] A tracer disturbance is injected into the vortex core origin point to identify the circulation path. At the determined vortex core origin point, a massless tracer disturbance is injected by numerical method, and the initial position of the disturbance is set as the origin point coordinates. The tracer disturbance moves in the flow field according to the Lagrangian method, and the motion equation is dx / dt=v(x,t), where x is the disturbance position vector, t is the time, and v is the flow field velocity vector. A high-precision interpolation method is used to interpolate the flow velocity value at the tracer disturbance position from the grid nodes, ensuring the continuity and accuracy of the disturbance motion. Multiple tracer disturbances are arranged in a concentric circle around the vortex core origin point, with the circle radius increasing outward from the vortex core center and the disturbances uniformly distributed. A dense disturbance cloud is formed. The motion trajectories of all tracer disturbances are tracked, and the trajectory set describes the complete circulation path structure and flow pattern of the vortex. The disturbance density analysis method is used to identify the circulation path boundary, which is defined as the position where the curvature of the disturbance trajectory changes sharply. The geometric features of the tracer disturbance trajectory are calculated, including trajectory length, average curvature, maximum deflection angle, and turning radius. By analyzing the topological structure of the disturbance trajectory, closed circulation paths and open streamlines are identified, and the bound flow inside the vortex and the through flow outside the vortex are distinguished.

[0036] The rotational parameters along the circulation path are extracted and encoded as vortex flow features. Based on the flow trajectory determined by the circulation path, the rotational motion parameters and flow features of each point on the path are calculated. The tangential velocity on the circulation path is calculated, which reflects the rotational speed of the fluid around the vortex core. The angular velocity is calculated, which describes the speed of the rotational motion along the circulation path. The circulation intensity Γ = ∮ v · dl is calculated, where v is the velocity vector, dl is the path element, and ∮ represents the line integral along the circulation path. This parameter quantifies the total rotational intensity of the vortex. The vortex radius is determined, which is defined as the radial distance at which the circulation intensity drops to a certain proportion of the maximum value. The strain rate parameter is calculated, which describes the deformation characteristics and shear intensity of the flow field on the circulation path. The vorticity magnitude is extracted, which represents the strength and concentration of the rotational motion. The vortex flow feature vector is established, which includes key parameters such as vorticity, circulation intensity, vortex radius, maximum tangential velocity, and strain rate. Principal component analysis is used to reduce the dimensionality of the feature vector, extract the most important vortex motion patterns, reduce the feature dimension, and highlight the main features. The standardized parameters after dimensionality reduction are encoded, generating a standardized vortex flow feature vector, and each component of the vector corresponds to different vortex motion characteristics.

[0037] Flow field disturbance coefficients are extracted from sea state dynamic parameters. Wind shear intensity is calculated from wind speed and direction data extracted from collected sea state dynamic parameters. Wind shear reflects the changing gradient of wind speed at different heights, directly affecting the vertical diffusion and mixing process of exhaust gas. The momentum exchange coefficient at the sea surface is calculated using the sea surface roughness theoretical model, establishing a quantitative relationship between the sea surface roughness length and the significant wave height in the sea state dynamic parameters z0 = αHs² / g, where z0 is the sea surface roughness length, α is the proportionality coefficient, Hs is the significant wave height, and g is the acceleration of gravity. This relationship describes the influence of sea waves on the airflow near the sea surface. The characteristic frequency of sea surface disturbance is calculated based on the wave period and wave height in the sea state dynamic parameters, which controls the periodic disturbance pattern and turbulence generation mechanism of the atmosphere near the sea surface. The difference between sea water temperature and atmospheric temperature in the sea state parameters is analyzed, and the buoyancy effect caused by the sea-air temperature difference is calculated, which affects the vertical transport and diffusion height of exhaust gas. A multi-dimensional flow field disturbance coefficient matrix is established by integrating the wind field, wave, and temperature field information in the sea state dynamic parameters. The coefficient matrix includes three main components: vertical disturbance coefficient, horizontal disturbance coefficient, and rotational disturbance coefficient. The vertical disturbance coefficient reflects the vertical turbulence intensity and mixing capacity, the horizontal disturbance coefficient describes the horizontal turbulence diffusion characteristics, and the rotational disturbance coefficient quantifies the environmental conditions for vortex generation and maintenance.

[0038] The vortex flow features are time-series correlated with the flow field disturbance coefficients to construct the exhaust gas-environment interaction field. The vortex flow feature vector and the flow field disturbance coefficient vector are aligned and matched according to the time sequence to establish synchronous feature-disturbance data pairs. The time delay between the feature vector and the disturbance coefficient vector is calculated through cross-correlation analysis to determine the optimal correlation time window. An interaction intensity matrix is constructed where F is the vortex flow feature vector, and D is the flow field disturbance coefficient vector, represents the tensor outer product operation. The matrix element I(i, j) represents the coupling strength between the ith vortex feature and the jth disturbance coefficient. The interaction strength matrix is extended in the time dimension to form a three-dimensional exhaust-environment interaction field tensor I(x, y, t), where x and y are spatial coordinates, and t is the time coordinate. The value at each spatial point in the interaction field represents the coupling degree of the exhaust vortex and the environmental disturbance at that location, and the higher the value, the stronger the interaction. The key influence area and sensitive period of exhaust diffusion are identified through the exhaust-environment interaction field, and the regulatory effect of environmental conditions on exhaust transmission is quantified.

[0039] In step S120, a high-temperature pollution source is located based on the exhaust-environment interaction field, a temperature field diffusion analysis is performed on the high-temperature pollution source to construct a heat propagation path, temperature gradient data are collected along the heat propagation path, and a waste heat flow network is mapped using the temperature gradient data.

[0040] In some embodiments, the locating of the high-temperature pollution source based on the exhaust-environment interaction field includes: extracting a thermal halo diffusion sub from the exhaust-environment interaction field; applying a reverse temperature gradient to the thermal halo diffusion sub to obtain a convergent trajectory; performing a thermal energy focusing analysis on the convergent trajectory to identify a singular point position; and determining the high-temperature pollution source based on the singular point position.

[0041] A thermal halo diffusion sub is extracted from the exhaust-environment interaction field. In the exhaust-environment interaction field tensor I(x, y, t), a spatial distribution pattern of temperature anomalies is identified, and the multi-scale features of the interaction field are decomposed by wavelet transform. The thermal halo diffusion sub is defined as a connected region in the interaction field where the temperature gradient modulus exceeds a threshold value, and the threshold value is determined by statistical analysis as the mean gradient plus twice the standard deviation. The identified temperature anomaly region is denoised using morphological opening operation to eliminate isolated noise points and small false targets. The processed region is analyzed for connectivity, and spatially continuous abnormal regions are marked as a single thermal halo diffusion sub. The geometric features of each diffusion sub, including area, perimeter, major axis direction, and shape factor, are calculated. An intensity distribution model of the thermal halo diffusion sub is established, and a Gaussian function is used to fit the temperature distribution inside the diffusion sub. The center position of the diffusion sub is defined as the temperature weighted barycenter, and the weight is the temperature increment relative to the background temperature. The time evolution information of the thermal halo diffusion sub is recorded to track the generation, development, and dissipation process of the diffusion sub. A diffusion sub database is established to store the characteristic parameters and evolution history of each diffusion sub.

[0042] A convergent trajectory is obtained by applying a reverse temperature gradient to the thermal halo diffusion sub. A reverse temperature gradient field is applied to the identified thermal halo diffusion sub region, and the gradient direction is opposite to the original temperature gradient. The reverse propagation equation is established as where a is the propagation coefficient, To diffuse the temperature gradient at the diffuser. The starting points are evenly distributed on the boundary of the diffuser. The number of starting points is determined by the size and complexity of the diffuser. The fourth-order Runge-Kutta method is used to solve the backward propagation equation and track the trajectory of the particle in the reverse gradient field. The calculation of the convergent trajectory uses adaptive step control, which reduces the step size in areas with large gradients and increases the step size in areas with flat gradients. Record the spatial coordinate sequence and corresponding temperature value of each trajectory, and establish the geometric description of the trajectory. Analyze the spatial relationship of multiple convergent trajectories, identify the intersection and bifurcation points of the trajectories. Calculate the curvature and torsion of the trajectory to quantify the geometric complexity of the trajectory. The end point of the convergent trajectory corresponds to the extreme position of the temperature gradient, and these positions are potential heat source candidates.

[0043] Perform heat energy focusing analysis on the convergent trajectory to identify singular point positions. Calculate the heat energy density distribution along the convergent trajectory, heat energy density ε = ρcpT, where ρ is the density, cp is the specific heat capacity, and T is the temperature. Use the sliding window technique to analyze the local variation characteristics of heat energy density on the trajectory, and the window size is determined according to the characteristic length of the trajectory. Calculate the second derivative of heat energy density d²ε / ds², where s is the arc length coordinate along the convergent trajectory. When the second derivative appears maximum, the corresponding position is marked as heat energy focusing point. Perform spatial clustering analysis on the focusing points of multiple convergent trajectories, and use the DBSCAN algorithm to identify the spatial clusters of focusing points. The singular point position is defined as the geometric center of the spatial region with the highest focusing point density. Calculate the heat energy intensity of the singular point position, which is equal to the total heat energy in the unit volume around the position. Analyze the stability of the singular point position, and verify the position change of the singular point under small perturbation through perturbation analysis. Establish a confidence evaluation system for the singular point, considering the number of focusing points, spatial concentration and heat energy intensity.

[0044] For example, determining a high-temperature pollution source based on the singular point position includes: evaluating the tracking accuracy based on the singular point position, the tracking accuracy including heat diffusion rate, environmental disturbance intensity and temperature gradient curvature; setting backtracking propagation parameters according to the tracking accuracy; and performing time sequence mapping on the backtracking propagation parameters to generate a high-temperature pollution source.

[0045] The tracking accuracy is determined based on the singularity position evaluation and complexity determination. A local coordinate system is established around the identified singularity position, and the thermal diffusivity α = k / (ρcp) of the region is calculated, where k is the thermal conductivity, ρ is the density, and cp is the specific heat capacity. The spatial variation gradient of the thermal diffusivity is calculated by numerical differentiation method, and the anisotropy characteristics of the thermal diffusion are identified. The environmental disturbance intensity at the singularity position is evaluated, and the disturbance degree is quantified by analyzing the standard deviation of wind speed fluctuation, pressure fluctuation, and temperature oscillation. A disturbance frequency spectrum analysis model is established to identify the main disturbance frequency components and energy distribution. The temperature gradient curvature κ = |d²T / dn²| is calculated, where T is the temperature and n is the direction perpendicular to the isotherm, and the curvature reflects the nonlinearity and complexity of the temperature field. The directional variation of the gradient curvature is analyzed, and the distribution characteristics of the principal curvature and secondary curvature are identified. A complexity comprehensive index C = w1·α + w2·σ_env + w3·κ is established, where w1, w2, and w3 are weight coefficients, and σ_env is the environmental disturbance intensity. The weight coefficients are determined by historical data statistical analysis, reflecting the influence degree of different factors on the tracing difficulty. The tracking accuracy requirement is determined according to the complexity level, and a nonlinear mapping relationship between complexity and accuracy is established. The accuracy level division standard is established, and the tracking accuracy is divided into four levels: rough, medium, fine, and super-fine.

[0046] The backtracking propagation parameters are set according to the tracking accuracy. Based on the determined tracking accuracy level, a hierarchical parameter configuration system is established, and different accuracy levels correspond to different backtracking propagation parameter combination schemes. The spatial step Δx is determined according to the spatial resolution requirement of the tracking accuracy, and the fine tracking requires smaller spatial step to capture subtle temperature changes. The time step Δt is set in association with the thermal diffusion characteristic time, and the higher the time resolution corresponds to smaller time step. The convergence criterion ε_conv is set according to the tolerance requirement of the tracking accuracy, and the high-precision tracking requires more stringent convergence standard to ensure the reliability of the calculation results. The interpolation order is selected according to the tracking accuracy requirement, and the linear interpolation is used for low-precision tracking, and the high-order interpolation method is used for high-precision tracking to reduce numerical errors. The boundary condition processing method is adjusted according to the tracking accuracy, and the fine tracking requires more accurate boundary processing technology and multiple point constraint conditions. The backtracking propagation parameters also include the maximum iteration number limit, numerical stability coefficient, and error control parameters, etc. The parameter sensitivity analysis model is established to evaluate the influence degree of different parameters on the accuracy of the backtracking results.

[0047] Temporal mapping of backtracking parameters generates high-temperature pollution sources. The set of backtracking parameters is segmented and mapped on the time axis, with each time segment corresponding to a different combination of parameter configurations. A temporal mapping matrix M(t) is established, with rows representing different types of backtracking parameters, columns representing time steps, and elements being the values of corresponding parameters at specific times. The temporal mapping process first discretizes continuous time into several time windows, with parameters remaining relatively stable within each window. At the boundaries of the windows, a smoothing transition function is used to avoid parameter discontinuity, with a cubic spline interpolation ensuring continuity. The mapping rules consider the time-varying characteristics of environmental conditions, including wind speed changes, temperature fluctuations, and turbulence intensity evolution. The mapped parameter sequence is input into the backtracking algorithm, which starts from the singular point location and performs time-reversal calculations. During the backtracking process, parameters are dynamically adjusted over time, with spatial steps automatically reduced in high-gradient areas and appropriately increased in low-gradient areas. Time steps are adaptively changed based on local thermal diffusion rates, with smaller time steps used in areas with faster diffusion. Through the control of temporally mapped parameters, the backtracking trajectory can accurately track the historical propagation path of heat. When all backtracking trajectories converge to a common spatial region, this region is determined to be the high-temperature pollution source. The determination process uses a multi-trajectory voting mechanism, with regions having more convergent trajectories having higher weights. The boundary of the final high-temperature pollution source is determined by the trajectory density contour, with the density threshold set according to the confidence requirement.

[0048] Temperature field diffusion analysis of high-temperature pollution sources constructs heat propagation paths. Based on the located high-temperature pollution source, a temperature field diffusion model is established to analyze the propagation process of heat from the pollution source to the surrounding environment. A three-dimensional temperature diffusion equation is constructed where α is the thermal diffusion coefficient and S is the high-temperature pollution source term. The heat intensity distribution of the pollution source is input as the source term, and the spatiotemporal evolution of the temperature field is solved. The finite element method is used to numerically solve the diffusion equation, with grid division considering the complex geometric boundaries around the high-temperature pollution source. The temperature gradient field is calculated, with the gradient direction indicating the main direction of heat propagation. Heat propagation paths are drawn along the temperature gradient direction using streamline tracing technology, with the paths starting at the center of the high-temperature pollution source and ending at the location where the temperature drops to the ambient temperature. The density of the heat propagation paths reflects the intensity of heat transfer in different directions, with higher-density regions corresponding to the main heat transfer channels. The path network topology is established to identify the main propagation paths and secondary branch paths. The heat propagation paths are classified into convection-dominated paths, conduction-dominated paths, and radiation-dominated paths, with significant differences in heat transfer mechanisms and propagation characteristics between different types of paths.

[0049] Temperature gradient data is collected along the heat propagation paths. Virtual temperature measurement points are arranged on the constructed heat propagation paths, with the measurement point spacing dynamically adjusted according to the path curvature and temperature variation rate. The temperature gradient components where s is the arc length coordinate along the path. The temperature gradient data includes parameters such as gradient magnitude, direction angle, and rate of change. The key nodes of the heat propagation path include path bifurcation points, curvature extreme points, and gradient mutation points. A spatial interpolation model of the temperature gradient is established, and the radial function interpolation method is used to calculate the gradient value in the inter-path region. The collected temperature gradient data is subjected to quality control to eliminate abnormal values and noise interference. The gradient data is classified and stored according to the path identifier, and each heat propagation path corresponds to a complete gradient sequence. The statistical characteristics of the temperature gradient data are analyzed, including mean, variance, kurtosis, and correlation length. The time sequence characteristics of the gradient data are established to track the evolution law of the gradient over time. The gradient data of different heat propagation paths are compared and analyzed to identify the similarities and differences between the paths.

[0050] The temperature gradient data is used to map the exhaust heat flow network. The heat flow intensity distribution is constructed based on the gradient magnitude in the collected temperature gradient data, and the heat flow intensity where k is the thermal conductivity of the material, is the temperature gradient. The spatial orientation of the heat flow vector is determined using the direction angle information in the temperature gradient data, and the convergence points and dispersion points of the heat flow are identified as the node positions of the exhaust heat flow network by combining the gradient change rate. According to the mean and variance distribution in the statistical characteristics of the gradient data, a weight distribution mechanism for network connection is established, and the main transmission channels of the network correspond to regions with higher gradient mean. The effective connection distance between network nodes is determined using the correlation length parameter of the temperature gradient data, and nodes beyond the correlation length are not directly connected. Based on the time sequence characteristics of the gradient data, the dynamic evolution law of the heat flow network is established to track the change process of the network topology over time. The parallel transmission paths and serial transmission paths in the network are identified using the similarity and difference analysis results between the gradient data of different heat propagation paths. The concentration degree of the heat flow is identified by the kurtosis parameter of the temperature gradient data, and the key transmission nodes in the exhaust heat flow network correspond to regions with higher kurtosis.

[0051] In step S130, the exhaust heat flow network is subjected to turbulent flow analysis to identify vortex breaking points, the mixing intensity parameters of the vortex breaking points are extracted, and the best treatment position is determined according to the correlation between the mixing intensity parameters and the vortex flow characteristics. The best treatment position is connected to the high-temperature pollution source to calibrate the treatment axis.

[0052] In some embodiments, the turbulent flow analysis of the exhaust heat flow network to identify vortex breaking points includes boundary scanning of the exhaust heat flow network to extract unstable regions, identification of energy dissipation peak points from the unstable regions, generation of a breaking index based on vortex reconstruction difficulty assessment of the peak points, and labeling of positions with a breaking index exceeding a threshold value as vortex breaking points.

[0053] Boundary scanning is performed to extract the instability region. A systematic scanning mechanism is established in the boundary region of the exhaust gas heat flow network to identify the flow instability phenomenon on the network boundary. The boundary scanning adopts the sliding window technique, and the window moves along the boundary curve of the heat flow network to monitor the flow field parameter changes near the boundary in real time. The boundary normal velocity gradient is calculated during the scanning process where v is the velocity amplitude and n is the boundary normal coordinate. When the normal gradient exceeds the stability threshold, the boundary segment is marked as a potential instability region. The disturbance growth characteristics of the network boundary are analyzed using linear stability theory to calculate the disturbance growth rate and the most unstable frequency. The boundary scanning also includes shear layer thickness measurement and mixing layer development analysis, and the sharp growth of the shear layer thickness indicates the occurrence of instability phenomenon. The geometric characteristics of the instability region are described, including the instability segment length, the maximum disturbance amplitude, and the disturbance propagation speed. The dominant instability mode in the instability region is identified through spectral analysis, distinguishing different mechanisms such as Kelvin-Helmholtz instability and Rayleigh-Taylor instability. The identification results of the instability region are marked on the heat flow network diagram through visualization technology, forming the instability distribution map.

[0054] Energy dissipation peak points are identified from the instability region. An energy dissipation rate calculation model is established in the extracted instability region to analyze the dissipation distribution characteristics of turbulent kinetic energy. The energy dissipation rate ε is calculated using the exact formula where ν is the kinematic viscosity coefficient, ui is the velocity component, and xj is the spatial coordinate. A high-density calculation grid is established in the instability region to ensure the capture of the spatial variation details of energy dissipation. The finite difference method is used to calculate the components of the velocity gradient tensor, and then the accurate dissipation rate distribution is obtained. The local extremum search algorithm is used to identify the energy dissipation peak points, which find the local maximum value position of the dissipation rate in the instability region. The peak point determination criteria require that the dissipation rate at this point not only has a local maximum value, but also exceeds a certain multiple of the average value. A strength classification system for peak points is established, and the peak points are divided into strong, medium, and weak levels according to the dissipation rate. The spatial aggregation characteristics of the energy dissipation peak points are analyzed to identify the clustering mode and distribution law of the peak points. The duration and stability of the peak points are calculated to exclude pseudo-peak values caused by transient fluctuations. The time evolution analysis of the peak points reveals the dynamic characteristics and periodic changes of the dissipation process.

[0055] The vortex reconstruction difficulty assessment is generated based on the peak point. A vortex reconstruction difficulty assessment system is established at the identified energy dissipation peak point position, quantifying the destruction degree of vortex structure. The reconstruction difficulty assessment considers multiple factors such as the continuity of vorticity field, the smoothness of velocity field, and the gradient distribution of pressure field. The vorticity loss rate around the peak point is calculated, which reflects the decay rate and irreversible destruction degree of vortex structure. A geometric complexity index is established to quantify the reconstruction difficulty by analyzing the twisting degree and topological changes of vortex lines near the peak point. The local properties of the velocity field at the peak point are analyzed using Laplace eigenvalues, and the distribution characteristics of the eigenvalues reflect the stability and reconstructability of the flow field. The calculation of the fragmentation index considers four dimensions: energy dissipation intensity, vorticity loss rate, geometric complexity, and topological changes. A normalized model of the fragmentation index BI=(E_diss / E_ref)×(Ω_loss / Ω_ref)×GC×TC is established, where E_diss is the dissipated energy, Ω_loss is the vorticity loss, GC is the geometric complexity, and TC is the topological change factor. The distribution characteristics of the fragmentation index of each peak point are determined through statistical analysis, and the probability density function of the index is established. The physical meaning of the fragmentation index represents the difficulty of restoring the vortex structure to its original state, and a higher index indicates more severe fragmentation.

[0056] The positions with fragmentation index exceeding the threshold value are marked as vortex fragmentation points. A fragmentation index threshold value determination method is established to determine a reasonable threshold value range through historical data statistical analysis and physical mechanism research. The threshold value is set using the quantile method, and the fragmentation index is sorted by size, with the upper quantile selected as the critical value for fragmentation determination. The ROC curve is used to analyze the accuracy and false positive rate of the fragmentation point identification under different threshold value settings, and the threshold value selection strategy is optimized. Spatial clustering analysis is performed on all peak points with fragmentation index exceeding the threshold value, and spatially adjacent fragmentation points are merged into fragmentation regions. The final marking of the vortex fragmentation point uses the weighted center method, and the weight of each point in the fragmentation region is determined by its fragmentation index. A confidence assessment system for fragmentation points is established, which considers the size, duration, and spatial consistency of the fragmentation index. The marking results of the vortex fragmentation point are visualized on the exhaust heat flow network diagram, and different colors and symbols are used to represent different intensity of the fragmentation point. A fragmentation point database is established to record the position coordinates, fragmentation index, generation time, and influence range of each vortex fragmentation point.

[0057] The mixing intensity parameters of the vortex breaking point are extracted. A local analysis domain is established at the identified vortex breaking point position, and the mixing characteristic parameters of the fluid in this region are calculated. The mixing intensity parameters include scalar gradient strength, strain rate amplitude, ratio of vorticity to strain rate, and turbulent mixing efficiency, among other key indicators. The scalar gradient strength reflects the degree of concentration field change at the breaking point, and the larger the gradient, the more intense the mixing. Eigenvalue analysis of the strain rate tensor reveals the main modes of fluid deformation at the vortex breaking point, including stretching, compression, and shear deformation. The turbulent mixing time scale τ_mix is calculated, which represents the length of time required to complete the mixing at the breaking point. A mixing efficiency evaluation model is established to quantify the mixing performance by comparing the actual mixing effect with the ideal mixing state. The extraction of mixing intensity parameters uses a multi-scale analysis method, taking into account the contributions of large-scale organizational motion and small-scale turbulent fluctuations. The probability density function is used to analyze the distribution characteristics of the scalar concentration at the vortex breaking point, identifying the mixing non-uniformity and intermittency characteristics. Time evolution analysis of the mixing intensity parameters reveals the dynamic characteristics and periodicity of the breaking process.

[0058] The optimal treatment location is determined based on the correlation between the mixing intensity parameters and the vortex flow characteristics. A multidimensional correlation analysis model is established between the mixing intensity parameters and the vortex flow characteristics to quantify the interaction strength between them. The Pearson correlation coefficient and the Spearman rank correlation coefficient are used to analyze the linear and nonlinear correlation between the parameters and the components of the characteristics. The correlation matrix R is constructed, where the matrix element R(i, j) represents the correlation strength between the i-th mixing intensity parameter and the j-th vortex flow characteristic. Principal component analysis is used to identify the parameter combination and characteristic combination that contribute most to the correlation. The spatial distribution of the correlation is established, and the spatial region with high correlation is identified in the exhaust heat flow network. An evaluation function F = w1 · R + w2 · A + w3 · E is designed for the optimal treatment location, where R is the correlation, A is the accessibility index, E is the treatment efficiency index, and w1, w2, w3 are weight coefficients. The accessibility index considers the connectivity and operability of the treatment location and the main channels of the heat flow network. The treatment efficiency index comprehensively evaluates the cost-benefit ratio and technical feasibility of treatment at this location. A multi-objective optimization algorithm is used to find the Pareto optimal solution in the candidate location, balancing the correlation, accessibility, and efficiency of the three objectives. Fuzzy evaluation method is used to determine the optimal treatment location, taking into account the uncertainty and fuzziness of multiple evaluation indicators.

[0059] The optimal treatment position is connected with the high-temperature pollution source to calibrate the treatment axis. A straight line is established between the determined optimal treatment position and the identified high-temperature pollution source, forming an initial geometric definition of the treatment axis. Considering the topological structure and flow resistance distribution of the waste heat flow network, the initial axis is path-optimized and adjusted. The path selection of the treatment axis follows the principle of the shortest resistance, avoiding high-resistance areas and dead angle areas in the heat flow network. The A* algorithm is used to search for the optimal path from the high-temperature pollution source to the optimal treatment position in the network, and the path cost function considers distance, resistance, and treatment effect. A three-dimensional space description of the axis is established, including the starting point, ending point, intermediate nodes, and curvature radius of the axis. The flow field characteristics along the axis are analyzed, and the positions of key control points and monitoring points on the axis are identified. The effective treatment length and influence radius of the axis are calculated to determine the arrangement spacing and coverage range of the treatment equipment. The calibration accuracy of the treatment axis is guaranteed by GPS positioning and laser ranging technology, with a calibration error controlled within centimeters. The treatment axis provides a spatial reference framework for the arrangement and operation of pollution control equipment.

[0060] In step S140, based on the treatment axis, the control parameter chain is reorganized, the control parameter chain is matched with the flow field disturbance coefficient to determine the steady-state treatment interval, and the adaptive adjustment matrix is generated using the steady-state treatment interval.

[0061] Specifically, the control parameter chain is reorganized based on the treatment axis. A control parameter reorganization framework is established along the calibrated treatment axis to systematically integrate the dispersed control elements according to the spatial order of the axis. The control parameters on the axis include key control variables such as flow regulating valve opening, temperature control set value, pressure regulating parameter, mixing ratio coefficient, and reaction residence time. The treatment axis is discretized into several control sections through spatial interpolation technology, and each control section corresponds to a specific parameter configuration. A priority order for parameter reorganization is established to determine the reorganization order based on the influence of each parameter on the treatment effect. The control parameter chain is constructed in a chain connection manner, with the output of the previous parameter serving as the input constraint condition for the next parameter. The interaction between each control point on the axis is analyzed to identify the synergistic and antagonistic effects between parameters.

[0062] In some embodiments, the matching of the control parameter chain with the flow field disturbance coefficient to determine the steady-state treatment interval includes: performing a resonance frequency scan on the control parameter chain to identify a resonance response mode; phase-locked to the resonance response mode and the flow field disturbance coefficient to obtain a synchronous coupling state; based on the synchronous coupling state, performing disturbance suppression analysis to determine a stabilization boundary; along the stabilization boundary, interval calibration is performed to generate a steady-state treatment interval.

[0063] The resonance frequency scanning is performed on the control parameter chain to identify the resonance response modes. Different frequency sinusoidal disturbance signals are injected into the control parameter chain one by one, and the frequency response characteristics of the control system are systematically scanned. The resonance frequency scanning starts from 0.1 Hz and increases to 100 Hz at logarithmic intervals, and each frequency point is kept for a sufficient time to allow the system to reach a steady-state response. The output amplitude and phase response of each link in the control parameter chain are recorded during the scanning process, and when the output amplitude shows a significant peak, it is marked as a resonance frequency point. A frequency response database of the parameter chain is established to store the gain and phase information of each control link at different frequencies. The resonance response modes are identified by analyzing the peak position and peak width in the amplitude-frequency characteristic curve, and each peak corresponds to a specific resonance mode. The quality factor Q value of each resonance mode is calculated, and the mode with a higher Q value indicates a sharper resonance, and the system response near the frequency is more sensitive. The dominant resonance mode and the secondary resonance mode in the control parameter chain are identified, and the dominant mode has the greatest impact on the system dynamic characteristics. The damping characteristics of the resonance modes are analyzed, and the damping ratio of each resonance response mode is determined by the half-power point method. A mode stability criterion is established, and the mode with a damping ratio greater than 0.1 is considered to be a stable resonance mode.

[0064] The resonance response modes are phase-locked with the flow field disturbance coefficients to obtain a synchronous coupling state. The phase information of the identified resonance response modes is extracted and matched with the phase distribution of the flow field disturbance coefficients. The phase locking process adjusts the phase compensator of the control system to keep the phase of the response mode synchronized with the main frequency component of the disturbance coefficient. A phase difference detector is established to monitor the phase deviation between the resonance response mode and the flow field disturbance coefficient in real time. When the phase difference remains within ±10 degrees for more than a preset time, it is determined that the phase locking is successfully established. The strength of the synchronous coupling state is quantified by the phase correlation coefficient, and a correlation coefficient greater than 0.8 indicates strong coupling, 0.5-0.8 indicates moderate coupling, and less than 0.5 indicates weak coupling. The locking ability of different resonance response modes and frequency components of the flow field disturbance coefficient is analyzed, and the mode with strong locking ability is preferentially selected for synchronization. A locking state maintenance mechanism is established to maintain the stability of the synchronous coupling state through continuous phase error feedback adjustment.

[0065] The stability boundary is determined by perturbation suppression analysis based on the established synchronous coupling state. Different intensities of test perturbation are injected into the system, and the perturbation suppression capability of the system is analyzed. The perturbation suppression analysis adopts an amplitude-increasing manner, starting from a small-amplitude perturbation and gradually increasing the perturbation intensity, and observing the response changes of the coupling state. The perturbation suppression ratio is calculated, which is defined as the ratio of the output perturbation amplitude to the input perturbation amplitude. The smaller the ratio, the stronger the suppression capability. The relationship curve between the perturbation intensity and the suppression ratio is established, and the critical perturbation intensity at which the system begins to lose suppression capability is identified. The stability boundary corresponds to the turning point at which the perturbation suppression ratio suddenly increases. After this point, the system's suppression capability for perturbation decreases sharply. The influence of the parameters of the synchronous coupling state on the position of the stability boundary is analyzed. Different coupling intensities correspond to different stability boundaries. The perturbation suppression characteristics at multiple frequency points are tested, and the frequency-dependent stability boundary distribution map is established. The most conservative boundary condition is determined, and the smallest stability boundary among the frequency points is selected as the overall stability boundary of the system.

[0066] The steady-state processing interval is generated by interval marking along the stability boundary. The determined stability boundary is used as a reference to demarcate the parameter interval within which the system can operate stably. The interval marking adopts a grid search method to systematically test the stability of each point within the boundary in the parameter space. Dense sampling is performed near the stability boundary to accurately determine the position and shape of the boundary. The boundary points of the steady-state processing interval are confirmed by stability testing. Only parameter combinations that pass the stability verification are included in the interval. The geometric description of the interval is established, and the vertex coordinates, boundary equations, and internal constraints of the steady-state processing interval are recorded. The effective area and volume of the interval are calculated to evaluate the size of the operable space for steady-state processing. The steady-state processing interval is managed by partitioning, and the interval is divided into high-stability, medium-stability, and edge-stability zones according to the stability level. The parameter configuration in the high-stability zone is preferentially recommended, and additional monitoring and protection measures are required when operating in the edge-stability zone.

[0067] An adaptive adjustment matrix is generated using a steady-state processing interval. Based on the determined steady-state processing interval, an adaptive adjustment matrix framework is established to convert the adjustment strategies within the interval into an operable mathematical form. The dimensions of the adaptive adjustment matrix are determined by the number of control points and the number of adjustment variables within the steady-state processing interval. The rows of the matrix correspond to the adjustment objects, and the columns correspond to the control inputs. The matrix elements A(i, j) = a(i, j) · b(i, j) · g(i, j), where a(i, j) is the basic gain coefficient determined by the static characteristics of the steady-state processing interval, b(i, j) is the adaptive coefficient assigned according to the interval stability classification (high stability zone, medium stability zone, and edge stability zone), and g(i, j) is the interval correction factor calculated from the effective area and boundary shape characteristics of the steady-state processing interval. The initialization of the matrix is configured using the spatial distribution of the control points within the interval and the coupling relationship of the adjustment variables. The adaptive mechanism realizes dynamic adjustment according to the partition management strategy of the steady-state processing interval. A larger adjustment gain is used in the high stability zone, and a smaller conservative gain is used in the edge stability zone. The sparsity structure of the adjustment matrix reflects the spatial adjacency relationship of the control variables within the interval, and a non-zero connection is established between adjacent control points. The constraint conditions of the matrix ensure that all adjustment operations are limited within the boundary range of the steady-state processing interval, preventing the system from deviating from the stable operating state.

[0068] In step S150, the adaptive adjustment matrix is fused with the temperature gradient data to identify the latent heat phase change region. The latent heat phase change region is extracted based on the condensation release energy and the energy difference between the waste heat flow network. The energy recovery tensor is constructed to generate a step-by-step processing scheme.

[0069] In some embodiments, the adaptive adjustment matrix is fused with the temperature gradient data to identify the latent heat phase change region, including: extracting the failure region distribution from the adaptive adjustment matrix; obtaining potential phase change points by difference fusion of the failure region distribution and the temperature gradient data; applying heat excitation to the potential phase change points to generate latent heat release response; determining the latent heat phase change region based on the latent heat release response.

[0070] Extract failure region distribution from adaptive tuning matrix. Identify regions with abnormal tuning gains in the adaptive tuning matrix, which correspond to locations where the control system response is abnormal or tuning is difficult. Failure regions are characterized by matrix element values deviating from the normal range, including abnormal cases such as excessive gain, insufficient gain, or gain oscillation. Establish failure criteria, and when the matrix element value exceeds the normal operating range, the location is marked as a failure point. Merge spatially adjacent failure points into continuous failure region distribution through connected component analysis. Analyze the spatial aggregation characteristics of failure regions to identify the distribution pattern and geometry of failure regions. Calculate the failure intensity of each failure region, which reflects the degree of tuning difficulty in that region. Failure region distribution usually corresponds to weak links and control blind spots in the system, which often have special physical phenomena. Establish a hierarchical system for failure regions, and according to the failure intensity, divide the regions into three levels: mild failure, moderate failure, and severe failure.

[0071] Differential fusion of failure region distribution and temperature gradient data to obtain potential phase transition points. Establish the spatial correspondence between failure region distribution and temperature gradient data, and ensure that both data sets are in the same coordinate system through spatial interpolation. Differential fusion uses a weighted difference method to calculate the difference distribution of failure region intensity and temperature gradient amplitude. Potential phase transition points correspond to the extreme positions in the fusion result, which represent regions where failure intensity and temperature gradient differ significantly. Analyze the statistical characteristics of the fusion result, including mean, variance, and skewness parameters, to identify outliers and outliers. The selection of phase transition points uses a local extreme search algorithm to find local maxima and minima in the fusion result. Establish a confidence evaluation system for potential phase transition points, considering factors such as fusion intensity, spatial consistency, and temporal stability. Perform spatial clustering analysis on phase transition points to merge spatially adjacent phase transition points into phase transition point groups.

[0072] Apply heat excitation to potential phase transition points to generate latent heat release response. Apply a control heat source at the identified potential phase transition point location to trigger the phase transition process at that location through heat excitation. Heat excitation uses pulse heating, and the heating power and duration are determined according to the characteristics of the phase transition point. Monitor the temperature response of the potential phase transition point during heat excitation and record the temperature-time curve. The latent heat release response is characterized by a plateau segment in the temperature curve, where the temperature remains relatively stable while heat is continuously input. Calculate the energy size of latent heat release by integrating the product of heat excitation power and platform duration to obtain the latent heat value. Analyze the latent heat release characteristics of different potential phase transition points to identify high-quality phase transition points with high latent heat release intensity and fast response time. Establish a classification system for latent heat release response based on latent heat size and release speed, and divide the response into strong response, medium response, and weak response. Record detailed parameters of latent heat release response, including latent heat value, release time, response delay, and decay characteristics.

[0073] The phase change latent heat region is determined based on the latent heat release response. The latent heat release response characteristics of all potential phase change points are comprehensively analyzed to identify the spatial region with significant latent heat release capacity. The boundary of the phase change latent heat region is determined by the isopleth of the latent heat release intensity, and a suitable intensity threshold is selected as the region boundary. The mass evaluation index of the latent heat region is established, including the total latent heat, release efficiency, spatial concentration, and time stability parameters. The phase change latent heat region is managed by partitioning, and the region is divided into high-efficiency, medium-efficiency, and low-efficiency zones according to the latent heat release characteristics. The spatial distribution of the latent heat region reflects the distribution of phase change materials and the spatial variation of phase change conditions in the system. The effective area and latent heat density of the phase change latent heat region are calculated to evaluate the energy recovery potential of the region. The determination of the phase change latent heat region provides a clear target area for subsequent energy extraction and recovery.

[0074] The condensation release energy is extracted in the phase change latent heat region. In the identified phase change latent heat region, the condensation process is triggered by cooling or pressurizing to convert gaseous substances into liquid and release latent heat. The extraction of condensation release energy uses a condenser array, which is arranged at key positions in the latent heat region to capture the energy release during the phase change process. The condensation temperature is controlled to be below the dew point temperature to ensure sufficient phase change and energy release. The intensity of energy extraction is controlled by adjusting the temperature and flow of the cooling medium, and the greater the cooling intensity, the more complete the condensation. The heat release during the condensation process is measured, and the value of the condensation release energy is calculated by enthalpy difference. The energy collection uses a heat exchange system to transfer the heat released during the condensation process to the working medium for recycling. A multi-point energy extraction network is established in the phase change latent heat region, and the condensation energy of each point is collected through a pipeline system for unified processing. The grade of the condensation release energy is represented by the temperature level, and the energy grade of high-temperature condensation release is higher and can be used for high-level applications. The energy extraction process needs to maintain suitable conditions in the phase change latent heat region to ensure the continuity and stability of the condensation process.

[0075] In some embodiments, the energy recovery tensor is constructed based on the energy difference between the condensation release energy and the waste heat flow network, including: obtaining a transient energy pulse by spatiotemporal decoupling of the condensation release energy; identifying an energy void by subtracting the energy pulse from the waste heat flow network; generating a recovery channel based on the potential energy of the energy void; and constructing an energy recovery tensor along the recovery channel by weaving a multi-dimensional mapping relationship.

[0076] The transient energy pulses are obtained by decoupling the temporal and spatial dimensions of the released condensation energy. The temporal and spatial dimensions of the extracted released condensation energy are analyzed to identify the timing characteristics and spatial distribution characteristics of energy release. The temporal and spatial decoupling uses singular value decomposition technology to decompose the energy space-time distribution matrix into the product of time mode and space mode. The time mode reflects the change law of the released condensation energy with time, and the space mode reflects the distribution mode of the energy in space. The transient energy pulses correspond to the peak part in the time mode, and these pulses represent the concentrated period of energy release. The time characteristics parameters such as the duration, peak intensity and repetition frequency of the energy pulse are analyzed. The envelope and carrier characteristics of the pulse are extracted to identify the modulation characteristics and spectral components of the transient energy pulse. A classification system of the pulse is established, and the energy pulse is divided into strong pulse, medium pulse and weak pulse according to the pulse intensity and duration. The spatial correspondence of the transient energy pulse is determined by the space mode, and the mapping relationship between the pulse and the spatial position is established. The energy density and power density of the pulse are calculated to evaluate the energy recovery value of the transient energy pulse.

[0077] The energy voids are identified by subtracting the energy pulse from the exhaust heat flow network. An energy comparison analysis framework of the transient energy pulse and the exhaust heat flow network is established to identify the energy difference and complementary relationship between them. The energy distribution of the energy pulse is subtracted from the energy distribution of the heat flow network point by point to obtain the spatial distribution of the energy difference. The energy void corresponds to the negative value region in the difference distribution, which indicates that the energy demand of the heat flow network exceeds the supply capacity of the energy pulse. The geometric characteristics of the voids are analyzed, including the area, depth, shape and distribution density of the voids. The intensity classification of the energy void is established, and the voids are divided into deep voids, medium voids and shallow voids according to the size of the energy gap. The spatial correlation between the voids is identified, and the merging and splitting of adjacent voids are analyzed. The total volume and average depth of the energy voids are calculated to evaluate the overall energy gap of the system. The distribution mode of the voids reflects the spatial characteristics and reasons of the energy supply-demand mismatch.

[0078] The recovery channel is generated by reconstructing potential energy based on energy voids. The identified energy void positions are used as target points for energy delivery, and a heat medium pipeline system is designed to connect the condensation energy release area and the energy demand area. The recovery channel is an actual pipeline network, including main pipelines, branch pipelines, and distribution pipelines, which are used to transport the heat released during condensation to the corresponding heat demand points of the energy voids. The main pipeline is drawn from the condenser outlet of the latent heat region and undertakes the main heat transfer task. The branch pipeline is designed according to the distribution of energy voids, with each void position corresponding to a branch pipeline connection. The distribution pipeline is set at the end of the branch pipeline and controls the heat supply to each energy void through the regulating valve. The pipe diameter of the pipeline is determined according to the depth and area of the energy void, and larger pipe diameters are used in areas with larger energy gaps. The working medium in the recovery channel is hot water or steam, which transports the heat released during condensation from the source to the demand point. Insulation measures and flow regulating devices are established for the pipeline to minimize energy loss during transmission. The layout of the recovery channel adopts a ring or tree structure to ensure the reliability and flexibility of the system's heat supply.

[0079] An energy recovery tensor is constructed by weaving a multi-dimensional mapping relationship along the recovery channel. Based on the designed recovery channel, a multi-dimensional energy transmission mapping relationship is established to represent the complex energy recovery process as a tensor. The multi-dimensional mapping relationship includes spatial mapping, temporal mapping, and energy mapping, which describe the transmission path of energy in space, the transmission process in time, and the conversion relationship of energy forms, respectively. The first order of the energy recovery tensor corresponds to the spatial position of the recovery channel, the second order corresponds to the time node, and the third order corresponds to the energy type. The tensor element T(i, j, k) represents the recovery intensity at the i-th spatial position, the j-th time node, and the k-th energy type. The weaving process converts the connection relationship of the recovery channel into the non-zero element pattern of the tensor, and the structure of the sparse tensor reflects the topological characteristics of the channel. Symmetry and positive definiteness constraints are established for the tensor to ensure the mathematical correctness and physical reasonableness of the energy recovery tensor. The main eigenvalues and eigenvectors of the tensor are calculated to identify the dominant and secondary modes of energy recovery. The compressed representation of the energy recovery tensor reduces the storage space and computational complexity.

[0080] A cascade processing scheme is generated based on the energy recovery tensor. Based on the energy distribution characteristics in the energy recovery tensor, a hierarchical processing flow is designed according to the energy grade from high to low. The cascade processing scheme corresponds to high-grade energy processing in the high-value area of the tensor, medium-grade energy processing in the medium-value area, and low-grade energy processing in the low-value area. The first stage processing targets the maximum value area in the tensor, and uses high-efficiency heat engines or power generation equipment for energy conversion. The second stage processing targets the medium-value area, and uses heating or process heating equipment for energy utilization. The third stage processing targets the smaller value area, and uses preheating or auxiliary heating equipment for energy recovery. The cascade connection mode of the scheme ensures that the waste heat of the upper level processing becomes the heat source of the lower level processing, realizing the step-by-step utilization of energy. The device selection of cascade processing is determined according to the energy size and temperature level in each area of the tensor, and the appropriate processing equipment is matched. The energy flow path of the scheme is established, and the energy transmission pipeline is designed along the gradient direction of the energy recovery tensor. The number of levels of the cascade processing scheme is determined by the numerical distribution characteristics of the tensor, and usually 3-4 processing levels are set to achieve the best energy utilization effect.

[0081] In step S160, the execution deviation of the cascade processing scheme is monitored to form a feedback curve, and a waste gas treatment execution instruction is generated based on the feedback curve to complete real-time processing control of marine waste gas.

[0082] In some embodiments, the monitoring of the execution deviation of the cascade processing scheme forms a feedback curve, which includes: extracting an inter-stage energy transfer rate from the cascade processing scheme; performing perturbation injection testing on the energy transfer rate to obtain a sensitivity distribution; identifying a deviation amplification node based on the sensitivity distribution; and fitting the time sequence response of the deviation amplification node into a feedback curve.

[0083] The inter-stage energy transfer rate is extracted from the cascade processing scheme. In the running cascade processing scheme, the energy transfer efficiency between adjacent levels is measured to quantify the flow characteristics of energy between the cascade levels. The inter-stage energy transfer rate is defined as the ratio of the actual received energy of the lower level to the output energy of the upper level, reflecting the loss in the energy transfer process. By installing energy measurement devices between each level, the input and output values of energy are monitored in real time. The calculation formula of the transfer rate is η=E_out / E_in, where E_out is the energy received by the lower level, and E_in is the energy output by the upper level. The time variation characteristics of the energy transfer rate are analyzed to identify the fluctuation rules and stability of the transfer efficiency. A statistical model of the transfer rate is established to calculate statistical parameters such as mean, standard deviation, and coefficient of variation of the transfer rate between levels. The spatial distribution of the inter-stage energy transfer rate reflects the energy utilization effect at different positions in the cascade processing scheme. The difference between the designed transfer rate and the actual transfer rate is compared to identify weak links with low transfer efficiency.

[0084] The sensitivity distribution is obtained by perturbation injection test. Small controllable perturbations are injected into the cascade process scheme, and the response of energy transfer rate to the perturbation is observed. Different types of test signals are used in the perturbation injection test, such as step signal, pulse signal and sinusoidal signal. Perturbations are injected at key positions of each process unit, including flow perturbation, temperature perturbation and pressure perturbation. The change amplitude of energy transfer rate before and after the perturbation injection is measured, and the sensitivity coefficient of transfer rate to the perturbation is calculated. The sensitivity distribution reflects the sensitivity of different positions in the scheme to the perturbation, and the position with high sensitivity will also have a significant response to small perturbations. A sensitivity matrix is established, with rows corresponding to perturbation injection positions and columns corresponding to energy transfer rate measurement positions, and matrix elements are sensitivity values. The spatial pattern of the sensitivity distribution is analyzed to identify the region with concentrated sensitivity and the path of sensitivity propagation. The amplitude of the perturbation is controlled within the range that does not affect normal operation, ensuring the safety and effectiveness of the test.

[0085] The bias amplification node is identified based on the sensitivity distribution. In the sensitivity distribution, key positions with high sensitivity values and large influence range are found, which correspond to the bias amplification nodes of the system. The identification of the node uses a threshold discrimination method, and when the sensitivity exceeds the set threshold and affects multiple downstream units, it is marked as an amplification node. The propagation characteristics of the sensitivity distribution are analyzed to identify the amplification path and attenuation path of the perturbation in the cascade process scheme. Bias amplification nodes are usually located at key links of energy transfer, such as heat exchanger, flow divider and control valve positions. The importance of the node is sorted, and the amplification node is prioritized according to the sensitivity and influence range. The spatial distribution of the node reflects the network structure and amplification mechanism of the bias propagation in the scheme. The physical characteristics of the bias amplification node are analyzed, including device type, operating parameters and boundary conditions. The monitoring focus of the node is established, and the bias amplification node is implemented with enhanced monitoring and early warning measures.

[0086] The time series response of the bias amplification node is fitted as a feedback curve. The response data of the bias amplification node at different times is collected, and the time series of the node response is established. The time series response includes the time history of temperature change, pressure fluctuation, flow regulation and energy transfer at the node. A suitable fitting function is selected to fit the response data of the node, and common functions include exponential decay function and second-order underdamped response function. The fitting of the feedback curve uses the least squares method, and the curve parameters are determined by minimizing the fitting error. The characteristic parameters of the fitted curve are analyzed, including response time, overshoot, regulation time and steady-state error. The slope change of the feedback curve reflects the response speed and stability of the bias amplification node. A family of feedback curves of multiple nodes is established, and the response characteristics and mutual relationship of different nodes are analyzed. The quality of the curve is evaluated by fitting goodness and residual analysis to ensure the accuracy and reliability of the fitting results.

[0087] The feedback curve is used to generate exhaust gas treatment execution instructions to complete real-time treatment control of marine exhaust gas. A decision mapping relationship between curve characteristics and control instructions is established, and threshold values are determined for the response time, overshoot, regulation time and steady-state error of the feedback curve. When the response time is less than 10 seconds, the response is marked as fast; when the overshoot is greater than 15%, the response is marked as excessive; when the regulation time is greater than 180 seconds, the regulation is marked as slow; and when the steady-state error is greater than 5%, the deviation is marked as out of limit. The control instruction priority is set, and the safety protection instruction has the highest priority. When multiple instructions conflict, the safety instruction is executed first and the other instructions are temporarily suspended. The response time of the feedback curve determines the starting time of the condenser. The node with a short response time starts the condenser immediately, and the node with a long response time starts the condenser later. The curve with a large overshoot triggers damping control measures, including reducing the inlet temperature of the heat exchanger, reducing the opening degree of the exhaust gas flow regulating valve, and starting the buffer tank to balance the pressure fluctuation. The regulation time parameter is used to determine the switching time of the cascade treatment scheme. When the regulation time is long, the current cascade configuration is maintained; when the regulation time is short, the next level of treatment is allowed to be switched quickly. When the steady-state error exceeds the threshold value, the compensation instruction is automatically started, the standby latent heat phase change area is opened, the sampling density of the energy recovery tensor is increased, and the gain coefficient of the adaptive adjustment matrix is adjusted. When the curve shows an exponential decay characteristic, the high-grade energy recovery device is gradually closed, and the low-grade energy treatment device is preferentially used. The second-order underdamped response curve starts the oscillation suppression measure, including closing part of the mixing device of the vortex breaking point, and reducing the response speed of the control parameter chain on the treatment axis. The node with a steep slope immediately cuts off the corresponding heat propagation path, starts the emergency cooling system and the exhaust gas bypass valve, and simultaneously activates the sound and light alarm device. A direct correspondence is established between the curve characteristics and the device instructions. The response time is mapped to the start delay, the overshoot is mapped to the damping strength, and the steady-state error is mapped to the compensation amplitude. The multi-node curve family is analyzed to realize the overall optimization control of the exhaust gas treatment system, and the safe and stable operation of the real-time treatment of marine exhaust gas is ensured.

[0088] In order to perform the intelligent control method of real-time treatment of marine exhaust gas corresponding to the above-mentioned method embodiment, the corresponding functions and technical effects are realized. Referring to Figure 2 , Figure 2 A structural block diagram of an intelligent control system 200 of real-time treatment of marine exhaust gas is shown. For ease of illustration, only the part related to the embodiment is shown. The intelligent control system 200 of real-time treatment of marine exhaust gas provided by the embodiment includes:

[0089] The data acquisition module 201 is configured to acquire a ship exhaust gas component concentration signal and a sea condition dynamic parameter, extract a vortex intensity from the exhaust gas component concentration signal to identify a vortex flow characteristic, extract a flow field disturbance coefficient from the sea condition dynamic parameter, and time sequence correlate the vortex flow characteristic and the flow field disturbance coefficient to construct an exhaust gas-environment interaction field.

[0090] The heat flow analysis module 202 is configured to locate a high-temperature pollution source through heat tracing based on the exhaust gas-environment interaction field, perform temperature field diffusion analysis on the high-temperature pollution source to construct a heat propagation path, collect temperature gradient data along the heat propagation path, and map an exhaust gas heat flow network by using the temperature gradient data;

[0091] The turbulent flow positioning module 203 is configured to analyze the exhaust gas heat flow network to identify a vortex breaking point, extract a mixing intensity parameter of the vortex breaking point, determine an optimal processing position according to a correlation degree between the mixing intensity parameter and the vortex flow feature, and label a processing axis by connecting the optimal processing position and the high-temperature pollution source;

[0092] The parameter configuration module 204 is configured to recombine a control parameter chain based on the processing axis, match the control parameter chain with the flow field disturbance coefficient to determine a steady-state processing interval, and generate an adaptive adjustment matrix by using the steady-state processing interval;

[0093] The energy recovery module 205 is configured to fuse the adaptive adjustment matrix and the temperature gradient data to identify a latent heat phase change region, extract condensation release energy in the latent heat phase change region, construct an energy recovery tensor based on an energy difference between the condensation release energy and the exhaust gas heat flow network, and generate a gradient processing scheme by using the energy recovery tensor;

[0094] The execution control module 206 is configured to monitor an execution deviation of the gradient processing scheme to form a feedback curve, generate an exhaust gas processing execution instruction based on the feedback curve, and complete real-time processing control of marine exhaust gas.

[0095] The above-described intelligent control system 200 for real-time processing of marine exhaust gas can implement an intelligent control method for real-time processing of marine exhaust gas according to the above-described method embodiments. The optional items in the above-described method embodiments are also applicable to this embodiment, and will not be described in detail herein. The remaining content of the present embodiment can be referred to the content of the above-described method embodiments, and will not be described in detail herein.

[0096] The purpose of the above-described embodiments is to exemplarily reproduce and deduce the technical solutions of the present application, and to completely describe the technical solutions, purposes and effects of the present application, so as to make the public more thoroughly and comprehensively understand the disclosed content of the present application, and not to limit the protection scope of the present application.

[0097] The above-described embodiments are not an exhaustive enumeration based on the present application, and there can be multiple other embodiments not listed. Any replacement and improvement made without violating the concept of the present application is within the protection scope of the present application.

Claims

1. An intelligent control method for real-time treatment of marine exhaust gas, characterized in that, The method comprises the following steps: Collecting ship exhaust composition concentration signal and sea state dynamic parameters, extracting vortex intensity from the exhaust composition concentration signal to identify vortex flow characteristics, extracting flow field disturbance coefficients from the sea state dynamic parameters, and constructing an exhaust-environment interaction field by time-series correlation of the vortex flow characteristics and the flow field disturbance coefficients; Based on the exhaust-environment interaction field, heat trace tracking is performed to locate high-temperature pollution sources, temperature field diffusion analysis is performed on the high-temperature pollution sources to construct a heat propagation path, temperature gradient data is collected along the heat propagation path, and the temperature gradient data is used to map the exhaust heat flow network; Performing turbulence analysis on the exhaust heat flow network to identify vortex breaking points, extracting mixing intensity parameters of the vortex breaking points, determining the optimal processing position according to the correlation degree of the mixing intensity parameters and the vortex flow characteristics, and connecting the optimal processing position with the high-temperature pollution source to calibrate the processing axis; Based on the processing axis, the control parameter chain is reorganized, the control parameter chain is matched with the flow field disturbance coefficient to determine the steady-state processing interval, and the steady-state processing interval is used to generate an adaptive adjustment matrix; Fusing the adaptive adjustment matrix with the temperature gradient data to identify the latent heat region of phase change, extracting the condensation release energy in the latent heat region of phase change, constructing an energy recovery tensor based on the energy difference between the condensation release energy and the exhaust heat flow network, and generating a cascade processing scheme through the energy recovery tensor; Monitoring the execution deviation of the cascade processing scheme to form a feedback curve, generating an exhaust treatment execution instruction based on the feedback curve, and completing real-time processing control of marine exhaust gas.

2. The method of claim 1, wherein, The vortex intensity extraction from the exhaust composition concentration signal to identify the vortex flow characteristics comprises: Constructing a vortex evolution graph based on the exhaust composition concentration signal; Performing reverse tracking processing on the vortex evolution graph to obtain a vortex core origin point; Injecting a tracer disturbance into the vortex core origin point to identify a circulation path; Extracting rotation parameters along the circulation path and encoding them as vortex flow characteristics.

3. The method of claim 1, wherein, The heat trace tracking based on the exhaust-environment interaction field to locate high-temperature pollution sources comprises: Extracting a thermal halo diffusion sub from the exhaust-environment interaction field; Applying a reverse temperature gradient to the thermal halo diffusion sub to obtain a convergence trajectory; Performing thermal energy focusing analysis on the convergence trajectory to identify a singular point position; Determining the high-temperature pollution source based on the singular point position.

4. The method of claim 1, wherein, The turbulence analysis on the exhaust heat flow network to identify vortex breaking points comprises: Performing boundary scanning on the exhaust heat flow network to extract an instability region; Identifying an energy dissipation peak point from the instability region; Performing vortex reconstruction difficulty evaluation based on the peak point to generate a breaking index; Labeling the position where the breaking index exceeds a threshold as a vortex breaking point.

5. The method of claim 1, wherein, The matching of the control parameter chain with the flow field disturbance coefficient to determine the steady-state processing interval comprises: Performing resonance frequency scanning on the control parameter chain to identify a resonance response mode; Phase-locked the resonance response mode with the flow field disturbance coefficient to obtain a synchronous coupling state; Performing disturbance suppression analysis based on the synchronous coupling state to determine a stabilization boundary; Performing interval labeling along the stabilization boundary to generate a steady-state processing interval.

6. The method of claim 1, wherein, The self-adaptive adjustment matrix is fused with the temperature gradient data to identify a phase change latent heat area, including: extracting a failure area distribution from the self-adaptive adjustment matrix; difference fusion of the failure area distribution and the temperature gradient data to obtain a potential phase change point; applying heat excitation to the potential phase change point to generate a latent heat release response; determining a phase change latent heat area based on the latent heat release response.

7. The method of claim 1, wherein, The energy recovery tensor is constructed based on the energy difference between the condensation release energy and the waste gas heat flow network, including: spatiotemporal decoupling of the condensation release energy to obtain a transient energy pulse; difference superposition of the energy pulse and the waste gas heat flow network to identify an energy void; potential energy reconstruction based on the energy void to generate a recovery channel; weaving a multi-dimensional mapping relationship along the recovery channel to construct an energy recovery tensor.

8. The method of claim 1, wherein, The execution deviation of the step-by-step processing scheme is monitored to form a feedback curve, including: extracting an inter-stage energy transfer rate from the step-by-step processing scheme; perturbation injection test of the energy transfer rate to obtain a sensitivity distribution; identifying a deviation amplification node based on the sensitivity distribution; fitting the timing response of the deviation amplification node into a feedback curve.

9. The method of claim 3, wherein, The high-temperature pollution source is determined based on the singularity position, including: evaluating the tracing complexity based on the singularity position to determine the tracking accuracy, the tracing complexity including the thermal diffusion rate, the environmental disturbance intensity, and the temperature gradient curvature; setting a backtracking propagation parameter according to the tracking accuracy; temporally mapping the backtracking propagation parameter to generate a high-temperature pollution source.

10. An intelligent control system for real-time treatment of marine exhaust gas, characterized in that, including: a data acquisition module, configured to acquire a ship waste gas composition concentration signal and a sea state dynamic parameter, extract a vortex intensity from the waste gas composition concentration signal to identify a vortex flow feature, extract a flow field disturbance coefficient from the sea state dynamic parameter, and time-correlate the vortex flow feature and the flow field disturbance coefficient to construct a waste gas-environment interaction field; a heat flow analysis module, configured to perform heat trace tracking based on the waste gas-environment interaction field to locate a high-temperature pollution source, perform temperature field diffusion analysis on the high-temperature pollution source to construct a heat propagation path, acquire temperature gradient data along the heat propagation path, and map a waste gas heat flow network using the temperature gradient data; a turbulent flow positioning module, configured to perform turbulent flow analysis on the waste gas heat flow network to identify a vortex breaking point, extract a mixing intensity parameter of the vortex breaking point, determine an optimal processing location according to the correlation degree between the mixing intensity parameter and the vortex flow feature, and connect the optimal processing location with the high-temperature pollution source to label a processing axis; a parameter configuration module, configured to reorganize a control parameter chain based on the processing axis, match the control parameter chain with the flow field disturbance coefficient to determine a steady-state processing interval, and generate a self-adaptive adjustment matrix using the steady-state processing interval; an energy recovery module, configured to fuse the self-adaptive adjustment matrix with the temperature gradient data to identify a phase change latent heat area, extract condensation release energy in the phase change latent heat area, construct an energy recovery tensor based on the energy difference between the condensation release energy and the waste gas heat flow network, and generate a step-by-step processing scheme through the energy recovery tensor; An execution control module is configured to monitor execution deviation of the stage treatment scheme to form a feedback curve, generate a waste gas treatment execution instruction based on the feedback curve, and complete real-time treatment control of the marine waste gas.

Images